Final Year Project Proposals

Semester 1, 2006

Electrical, Computer & Telecommunication Engineering

(There are Software Engineering Projects at the end of this document.)

Any student enrolled in this course can put any project offered as a preference. Please list
a minimum of 3 projects using the project code (eg COA-06-1), title (egHigh Performance
Control of a Synchronous Reluctance Motor Drive) and supervisor (eg Colin Coates). If you
provide any other information about your interest areas then this will also be taken into
account. Please refer to the Level of Difficulty of the project before you choose it as this
is a guide to indicate the difficulty of the expectation of the project outcomes.

Colin Coates

COA-06-1: High Performance Control of a Synchronous Reluctance Motor Drive

Proposed By: Colin Coates
Industry Partner: None
Description: A synchronous reluctance motor is an alternative to the induction motor in
variable speed drive applications. Recent research has indicated advantages to having
more than three phase windings on these machines. The primary advantages are that
the motor produces greater torque with reduced losses. An experimental nine-phase
machine and inverter have been built. The aim of this project is to implement a high
performance controller in the drive. This will require innovative use of a DSP controller
and XILINX FPGA along with the development of efficient algorithms. This project
covers areas in electric machines and drives, DSP, electronic design and control.
Outcomes: This project has been attempted previously and a student would be required
to build on prior results. Experience suggests there is a steep learning curve required
with this project. Before a student can accomplish anything they need to gain familiarity
with the Texas instruments TMS320C6711 DSP, Code Composer Studio development
environment, XILINX XCS40XL FPGA, XILINX ISE development environment and the
inverter hardware. The required outcome is to successfully negotiate this hump,
implement a working drive (moderate to difficult) that uses vector control techniques
(difficult to challenging) and undertake performance measurements on the drive.
Level of Difficulty: Moderate to Challenging
Required Background: ELEC3130 (ELEC2160, ELEC3100), ELEC3250
Number of Students: 1

COA-06-2 : Position Sensor Less Control a Switched Reluctance Motor

Proposed By: Colin Coates
Industry Partner: None
Description: This aim of this project is to modify a switched reluctance drive so that it
can function without a rotor position sensor. Previous project students have developed a
speed controller for a switched reluctance motor. However, this controller requires
knowledge of rotor position. This project seeks to use stator voltage and current
measurements to determine the rotor position avoiding the need for a rotor position
sensor.
Outcomes: The outcome (difficult) anticipated is a working position sensor less switched
reluctance drive.
Level of Difficulty: Difficult
Required Background: ELEC3130 (ELEC2160, ELEC3100)
Number of Students: 1

COA-06-3 : Integrated Rectifier / Boost Converter / Inverter for Renewable

Energy Applications
Proposed By: Colin Coates
Industry Partner: Aerogenesis
Description: The aim of this project is to adapt a standard ac drive circuit for use in low
power renewable energy applications that employ an induction or permanent magnet
generator (eg. wind, water turbines). The circuit will be required to accept a variable
(three phase input), produce a regulated DC bus voltage and regulated ac voltage
(single phase) output. The circuit should be capable of operating as a stand-alone
supply or connected to the supply mains. The project will require theoretical verification,
simulation, design and construction of a working prototype.
Outcomes: The outcome associated with a difficult level difficulty project is a working
prototype inverter.
Level of Difficulty: Difficult
Required Background: ELEC3250
Number of Students: 1

COA-06-4: Bi-directional DC/DC Converter

Proposed By: Colin Coates
Industry Partner: Aerogenesis
Description: The aim of this project is to develop a bi-directional dc/dc converter to
interface a battery bank to the DC bus of an inverter circuit for a renewably energy
source. The converter should allow for the supply of energy to and from battery storage
system as system conditions demand. The project will require theoretical verification,
simulation, design and construction of a working prototype.
Outcomes: The outcome associated with a difficult level difficulty project is a working
prototype bi-directional DC/DC converter.
Level of Difficulty: Difficult
Required Background: ELEC3250
Number of Students: 1

COA-06-5 : Simulation Study of Parallel inverters Operating in a Low Voltage

Distributed Generation System
Proposed By: Colin Coates
Industry Partner: CSIRO Energy Technology
Description: A new concept in power generation / distribution is that of a micro-grid. A
micro-grid is a situation where an industrial site or building might have several small
alternative electrical energy sources (solar, wind, gas turbines, batteries, fuel cells)
associated with it in addition to a connection to the main supply grid. The motivation for
this could be greater efficiency, reduced greenhouse gas emission or improved reliability
of supply. This project aims to specifically look at the modelling of the inverters that
make up these systems and their control strategies. Simulation studies will be carried
out to verify system behaviour to sudden load changes, unbalanced loads and other
likely scenarios.
Outcomes: Variants of this project have been undertaken previously and a student
would be required to build on prior results. The outcome associated with easy to
moderate level of difficulty would be to research and verify models for inverters and
known control strategies. An additional outcome associated with a difficult to challenging
level of difficulty is to generate methods to potentially improve the network behaviour of
a low voltage distributed generation system.
Level of Difficulty: Moderate to Challenging
Required Background: ELEC4100
Number of Students: 1

Jose De Dona
DeD-06-1 : A Ball-In-A-Tube Height Control Experiment
Proposed by: Dr. Jose De Dona
Description: The purpose of this project is to design and build a ball-in-a-tube
apparatus; and to design and implement a control algorithm to be able to control the
height of the ball. The ball-in-a-tube is a simple and inexpensive experimental setup that
can be used to demonstrate some of the fundamental concepts of feedback control
systems. It consists of a transparent tube in a vertical position. Inside the tube there is a
light weight ball (e.g., a ping-pong ball); at the bottom of the tube there is a fan that is
used to lift the ball; and at the top of the tube there is a sensor (e.g., an ultrasonic
sensor) to measure the ball's height. The measurements of the height sensor are sent to
a computer through an A/D acquisition card and the computer is programmed to
implement a control strategy to, e.g., keep the ball at a fixed desired height position or to
make the ball follow a desired trajectory. The resulting control signal is then output
through a D/A conversion card to the motor that drives the fan (this signal can be an
analog continuous signal or a PWM signal).
The control algorithm can be implemented in a software package such as ProcessACT,
dSPACE (which can be interfaced with Matlab/Simulink for controller design and
graphical and GUI capabilities), LabVIEW, Matlab/Simulink/Real Time Workshop, etc.
Another possibility is to construct an embedded control system consisting of a micro-
controller attached to the experimental set up, with a connection to a PC for controller
design/downloading and visualization.
Outcomes: The expected outcome is the complete construction of the experimental set
up, together with the required connection to the computer (or micro-controller) for
controlling the ball height, and the implementation of a control algorithm to demonstrate
the functionality of the equipment. It is expected that the design of the controller be
customisable from the computer, in order to be able to demonstrate different control
strategies and to be able to make the ball follow trajectories defined by the user. Also,
the program should have good graphical capabilities to generate plots of the time
response of variables of interest for reporting and comparison of different control
designs.
Skills Required &/or Developed: The skills necessary to build the apparatus are basic,
as it is a simple setup. Standard electronic skills could be required, to be able to design
the few circuitry (if any) that could be required to perform the electrical connections.
Basic standard programming skills may be required depending on the software platform
chosen to implement the control law (e.g., C language). Basic knowledge of
Matlab/Simulink is required to perform the system modelling and to design, evaluate and
validate different control designs. If the chosen design is based on a micro-controller,
then micro-controller programming skills would be required or could be developed during
the project. The essential skills that are needed for this project are basic system
modelling, signal processing and control system design. Hence, the student attempting
this project should have a solid knowledge of the concepts imparted in subjects such as
ELEC2400, ELEC3400 and ELEC4400.
Level of Difficulty: Construction of the apparatus: Easy
Design of the control laws and software implementation: Easy to Challenging, depending
on the features included and the control techniques implemented.
Number of Students: This is a project suited for one student. (It can be turned into a
project for more students by implementing a multiple-tube multiple-balls system sharing
a common input manifold inlet, to study advanced control concepts and resource
allocation strategies. This way it might become a quite challenging project on
multivariable control.)

DeD-06-2 : Autonomous Underwater Vehicle

Proposed by: Dr. Jose De Dona
Description: The purpose of this project is to design, construct and to demonstrate
operability of an Autonomous Underwater Vehicle (AUV).In recent years, underwater
robotics has become a hot area of research due to the availability of emerging enabling
technologies. AUV's are used today for undersea exploration, monitoring, rescue
operations and other scientific tasks, often in hostile conditions and without endangering
human lives. The project will consist of a number of parts, which will have to be well
integrated with each other for a successful outcome. Some of these parts are:
*Design and construction of the AUV's body. The design of the AUV's body is left open to
students, but as an initial suggestion it could consist of a primary hull of cylindrical shape
that will contain all the electronic components, and two secondary hulls, also of
cylindrical shape, attached underneath the primary hull that will contain the propulsion
system (dive system and maneuvering system), ballast and floating devices. In order to
prevent electrical discharges and leaks to the environment, the power source and all
electronic components will be enclosed in waterproof encasing.
* Design and implementation of the on-board instrumentation equipment. The AUV,
being unmanned, needs to sense a number of physical and position variables, in order
to be able to take decisions and react to the environment, and to accomplish its mission
successfully. A number of sensors is thus required. An initial suggestion of required
sensors is: a sonar sensor to detect obstacles, a colour camera for pattern and object
recognition, compass sensor to compute the vehicle's heading, pressure sensor to
compute the vehicle's depth, accelerometers, leak sensor to detect the presence of
water leakage, etc.
* Design of the control system and navigation algorithms. This will require the selection
of a micro controller and associated circuitry. The micro controller will combine the
environmental and vehicle position information provided by the sensors and will
implement a number of control algorithms to achieve the navigation and environmental
sensing objectives, as well as a number of safety procedures in case something goes
wrong.
Outcomes: The minimum outcomes that are expected for this project are:
* Design and construction of the vehicle,
* Instrumentation and control algorithms that allow the safe undertaking of simple
underwater navigation operations carried out in an autonomous way,
* Demonstration of the operation of the vehicle, safe unmanned operation, and basic
navigation manoeuvres.
* The constructed vehicle is intended to be used in subsequent projects as a prototype
for future designs and to be used in projects studying the coordination between several
such AUV's. Hence, it is expected that the construction be done at the lowest possible
cost and in such a way that future modifications and upgrades, and the inclusion in the
vehicle of extra equipment (e.g., communications equipment) be facilitated. It is also
expected that the features and operation instructions of the constructed vehicle be well
documented to facilitate future developments.
Skills Required &/or Developed: Some of the skills that will be needed are:
* Mechanical design: To construct the vehicle to comply with the requirements of
underwater operation, including drag reduction, determination of the correct amount of
buoyancy, resistance to water pressure and waterproofing.
* Electrical design: To implement the propulsion system including power supply and
electrical drivers.
* Electronic design: To construct the hardware circuitry, to integrate the different sensors
and interface them with the micro controller, and to operate the electrical drivers.
* Micro controller programming: To be able to implement the main control and data
acquisition algorithms.
* Signal processing: To be able to estimate different variables, including vehicle position,
from the sensed measurements.
* System dynamics and control: To develop models for the vehicle dynamics and to be
able to generate control algorithms to achieve vehicle stability, to perform navigation
operations, trajectory tracking, etc.
Level of Difficulty: Challenging
Number of Students: This is a project suited for two or three students. In the case of
a two students group, the project can be undertaken as described above. If there is a
third member in the group, then more additional features will be expected, e.g., the
implementation of more advanced (as compared to a two students project) model based
control techniques for trajectory tracking in the presence of disturbances, the execution
of more complicated manoeuvring operations, more advanced image processing
capabilities using the video camera, etc.

Minyue Fu

FU-06-1: Sony Training Projects

Supervisor: Minyue Fu
Description: The Japanese Electronics Company Sony has a selection of projects/job
that they have requested we advertise for academically competent final year students.
The approximate arrangement with each project is for the student to work on the project
in Australia in Semester 1 and then with Faculty funding spend between 1 and 3 months
in Japan at Sony finishing work on the project. Japanese language proficiency is not
needed excepted where specified. The following projects are on offer:
 1. Measurement, Fabrication and Analysis of electric characteristics of organic thin-
film transistor.
2. Measurement & Fabrication of electric characteristics of thin-film transistor
(Atsugi)
3. Sputtering multiplayer films using high performance sputtering machine.
Evaluation of magnetic thin film and some preliminary measurement such as X-
ray, Sheet resistance, Roughness.
4. Characterization of super high power red semiconductor array lasers to improve
the device characteristics and or fabrication of laser light source module and the
application to a display. (Some level of Japanese language preferred)
Required background:
Level of Difficulty: Challenging
Number of Students: 1 student per project

FU-06-2: Acoustic Beamforming using Ultrasonic Transmitters

Supervisor: Minyue Fu
Description: Acoustic beam forming has many applications including surround sound
systems. The basic idea is to use an array of speakers to create dynamic and realistic
sound effects. The key is the localization of acoustic signals. However, the degree of
localization is typically limited due to the simple fact that the acoustic wavelength is large
(about 68cm for 500Hz signals). In this project, the aim is to study the use of an array of
ultrasonic transmitters (speakers) for beam forming. Due to the much smaller
wavelength, ultrasonic signals are much more suitable for localizing signals. Due to the
natural nonlinear demodulation property of the air, it is possible to transmit acoustic
signals through the air using two correlated ultrasonic signals. The result will be a highly
localised acoustic "spotlight". A wide range of applications of the technique exists in
museums, public facilities or even vehicles where personalized acoustic spotlights are
ideal for information delivery. An example of it is that when you stop exactly in front of a
painting in a museum, you can hear the commentary for this particular painting.
Required background: Good understanding of Digital Signal Processing, competence
in mathematics, good skills in digital design and software development. ELEC3240 or
ELEC3210
Level of Difficulty: Moderate to Difficult
Number of Students: 1-2

FU-06-3: Dual-Stage High Precision Control

Supervisor: Minyue Fu
Description: Dual-stage control systems are popularly used high precision positioning
and tracking. It consists of two coupled control systems, a coarse control stage providing
long-range but low precision and low bandwidth control and a fine control stage
providing small-range but high precision and high bandwidth. A best biological example
is perhaps your hand where the arm acts as a coarse control stage and the fingers act
as fine control stages. Industrial applications of dual-stage control are virtually endless.
In principle, any given control system can be easily reconfigured by adding a fine control
stage to provide a much higher precision control. However, the strong decoupling
between the two stages make the control design task a very interesting exercise. The
purpose of this project is to design a simple two-stage high precision control system and
study the associated control design problems.
Required background: Good understanding of modeling, control and digital signal
processing, good hands-on skills in software and hardware development. ELEC4400
Level of Difficulty: Moderate to Difficult
Number of Students: 1-2

FU-06-4: Magnetic Levitation System

Supervisor: Minyue Fu
Description: This project aims to develop a set of "magnetic bearings" to replace the
conventional mechanical bearings. The magnetic bearings are used to support high
speed rotational machines. These bearings have several advantages: no mechanical
friction; no vibration noise; and less dissipation of heat. The project involves both
hardware and software designs. The hardware design involves mechanical, electrical
and electronic parts. Skills required or acquired: Basic knowledge of EE, CE and/or
Mechatronics.
Required Background: ELEC3240 or ELEC3210, ELEC2120 ELEC2130
Level of Difficulty: Moderate to Challenge.
Number of Students: 1-2

Jamil Khan
KHA-06-1: Design and Development of a PSTN to VoIP Handset Hardware Interface
Using a General Purpose Microprocessor
Description: The project needs to develop a hardware interface that will enable a PSTN
handset to be used as VoIP telephone using a LAN connection. The hardware need to
be developed using a general purpose microprocessor and other necessary ICs. The
hardware will convert all the signals and a voice signal from a PSTN phone to VoIP
format or VoIP format to analogue format in the opposite direction. A PSTN phone is a
central battery system which requires DC supplies and various tones for signalling
purposes. On the other hand a VoIP phone is a main powered phone which a use text
based signalling. The project need to develop necessary hardware and associated
software for the interface.
Additional Task: Design an Ethernet port for the microprocessor. Develop a single
interface board with a pic processor.
Outcomes: A working interface board that can connect a telephone handset to a PC
which can exchange voice in both directions and also able to exchange signals. The
interface should allow any standard telephone handset can be used as a VoIP
telephone.
Level of Difficulty: Moderate to Challenging (with all additional tasks)
No. of students: 1
Background Knowledge: ELEC2700, ELEC2500, ELEC3500.

KHA-06-2: Design and Development of a Wireless Remote Object Monitoring

System
Description: Main objective of this project is to develop a remote/network based locking
and monitoring system. The idea is to develop a polling network with a controller that
scans multiple objects and devices which will be electronically locked. The controller
periodically scans all the devices using a RF link and receives responses to confirm their
identity and possible locations/distance from the controller. The controller should be able
to generate an alarm if any object is moved from a particular position. Also, the controller
should be able to identify a missing object or objects. In order to remove an object from
its specified location a valid password must be used which will be authorized by the
controller. An authorized removal should not generate any alarm.
The project has two modes, basic and advanced. Basic mode is for a single student and
the advanced mode is for two students. The project is an open ended project.
Outcomes:
Following main tasks are involved in the basic project:
Design and develop hardware and software for remote devices and the controller.
Designed hardware will be based on a microprocessor.
Design and develop radio links.
Interface appropriate sensors if necessary.
Develop a method of identifying each objects and its distance from the controller.
Advanced task:
Study and implement a triangulation technique to precisely workout the location of a
device.
Develop a Bluetooth link to transmit video frames.
A camera interface on a selected object to transmit pictures from its surrounding.
Able to display transmitted video frame on a PC monitor.
Difficulty: Moderate to challenging.
Project type: Design, development and construction of hardware and software modules.
No. of students: 1 (basic), 2 (advanced).
Background Knowledge: ELEC2700, ELEC2500, ELEC3500.

KHA-06-3: Design and Development of a Car Collision Avoidance System

Description: The project will develop hardware and software for automobiles to develop
a collision avoidance system. The idea is to develop a car unit that transmits data using
a low power short range radio transmitter. This unit will transmit information such as card
ID, speed, direction of travel, acceleration data, etc. Other cars when comes within the
transmission range of the car the receivers should be able to read those data then
calculate the safe distance and speed, and display those information for the driver. The
unit should be able to pick up information from multiple cars and calculate the possibility
of a hit and instruct or advise the driver to change the speed or take other evasive
actions.
Outcomes: Develop a microprocessor unit a multi-line display.
Interface a radio transceiver with the microprocessor board.
Implement a packet transmission or on demand information packet transmission
algorithm.
Implement a multi-car collision avoidance system.
Test the algorithm in realtime.
Difficulty: Moderate to difficult.
Project type: Design, development and construction of hardware and software modules.
No. of students: 2.
Background Knowledge: ELEC2700, ELEC3500.

KHA-06-4: A Simple Solar Powered Remote Self Contained Unit to Measure Traffic
Volume on a Road
Description: The basic idea is to build a self contained solar powered unit to remotely
monitor traffic (vehicles) on a road. The remote unit will transmit traffic information to a
controller located at a certain distance. The unit should be able to measure traffic volume
(no. of vehicles/minute) as well as monitor traffic flows and detects congestions.
Appropriate sensors need to be interfaced to the self-contained board for monitoring
traffic. A solar cell based power supply will be designed for the system to power the
board. The remote board will be connected to a PC via a radio link. Appropriate software
need to be developed to display traffic information.
Outcomes: Design and construction of a solar power supply.
Interfacing sensors and RF/radio devices.
Development and implementation traffic monitoring algorithms.
Design and Development of a display software on a PC.
Multilane traffic monitoring algorithm (advanced).
Realtime video monitoring and car speed measurement (advanced).
Difficulty: Moderate to challenging.
Project type: Design, development and construction of hardware and software.
No. of students: 1 (basic) 2 (advanced).
Background Knowledge: ELEC2700, ELEC3240.

KHA-06-5: Design & Development of a Power Line Communication System

Proposed by: Jamil Khan (co-supervised by Dr. Colin Coates)
Description: Power line communication is one of the important growth areas of data
networking. This project will develop a power line data communication interface using
specialised power line communication integrated circuit. Full hardware and software
need to be developed to exchange data between different terminals connected to a
home power line. The transmission data rate of the interface must be at least several
hundred kbs. The developed hardware should be able to accept data from a PC to
transmit over the power line. A RS-232/USB connection will be used to connect the
developed interface with a PC.
Outcomes:
Fully understand the powerline communication techniques.
Design and develop a powerline data interface circuit to transmit data for home
applications.
Duplicate the developed interface to setup a testbed for data applications.
Interface the developed interface to transmit data from a PC.
Difficulty: Difficult to challenging.
Project type: Design, development and construction of hardware and software.
No. of students: 1.
Background Knowledge: ELEC2700, ELEC3240, ELEC3500.

KHA-06-6: A Simple Wireless Sensor Network for Environment Monitoring and

Control of Equipments
This project is concerned with the development of a multisensor based large building
monitoring system that can use realtime information to identify different physical events
such as fire, temperature, movement of people, etc. A multi-node wireless sensor
network needs to be developed using the zigbee standard hardware. The project needs
to develop the sensor interface circuits which can be connected to the zigbee hardware.
Zigbee hardware has already been developed in the electrical engineering laboratory.
Within the sensor network all the sensors should send their data to a central controller
for processing. The central controller will control an equipment based on the received
data. One of the applications could be an auditorium temperature control. Assume a
large auditorium where multimode sensors will be placed in four different corners to
monitor temperature and humidity. Based on the received data from multiple sensors the
central controller could control fans or air conditioners to maintain a uniform temperature
within the auditorium. This solution can be extended for other scenarios.
Outcomes:
Understand the zigbee standard.
Design and develop sensor interfaces.
Establish network connections between multiple nodes.
Develop algorithms for condition monitoring and equipment control.
Difficulty: Difficult to challenging.
Project type: Design, development and construction of hardware and software.
No. of students: 1.
Background Knowledge: ELEC2700, ELEC3500.

KHA-06-7: An OPNET simulation model to study a body area sensor network

Proposed by: Jamil Khan (co-supervised with Dr. Mehmet Yuce)
The project will develop a body sensor network simulation model to study the
performance of a body network for different medical application scenarios. This is a
software based project. OPNET will be used to develop traffic models, network models,
antenna models and application models. The model should also study the performance
different types of antennas. The project need to develop new models including
programming the OPNETs pipeline process.
Outcomes:
Understand the wireless sensor and body networking areas.
Fully understand OPNET modelling and programming techniques.
Develop body network architectures for different applications.
Develop OPNET based simulation models.
Study v.
Difficulty: Difficult to challenging.
Project type: Develop simulation models based on the OPNET package.
No. of students: 1.
Background Knowledge: ELEC2500, ELEC3500.

Abbas Kouzanis

KOU-06-1: Smart Undervehicle Inspection System

Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: Under vehicle inspection is currently performed by a person walking
around the vehicle with a convex mirror attached to the end of a stick. The inspection is
carried out at the security entrance of a government, military, or private facility through
what the person can see on the mirror to identify objects of security threats. The aim of
this project is to develop an automated system that would allow a complete
undercarriage inspection of a vehicle without requiring the use of the mirror-on-the-stick
approach. The system will use a 1D array of cameras as well as lighting installed in a
low-ramp housing over which vehicle drives, as well as a processing platform. The
processing platform will run an image mosaicing algorithm that will fuse images obtained
from the cameras to generate different perspective views that can be used to make a 3D
representation of a vehicle undercarriage. In addition, the processing platform will run an
image analysis algorithm to automatically identify suspicious objects placed under
vehicle. This project is suitable for a group of three students. The first student will be
responsible for the development of the 1D camera array, lighting, processing platform,
and interfacing. The second student will be responsible for the development of the image
mosaicing algorithm and necessary GUI. The third student will be responsible for the
development of the suspicious-object detection algorithm.
Outcome: A system including hardware and software that captures multiple 2D images
from an array of USB cameras, generates and displays 3D representation of the vehicle
undercarriage from the 2D images, and detects suspicious objects.
Level of Difficulty: Challenging
Required Background: Student 1: Good understanding of digital design techniques
including VHDL, microprocessor and microcontrollers, and low level hardware-related
programming. Students 2&3: Strong programming skills, good understanding of digital
signal processing techniques, and competence in mathematics.
Number of Students: up to 3

KOU-06-2: Portable 3D Face Recognition System Using Camera Array and

Reconfigurable Hardware
Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: The face is a distinct part of who the person is and how people recognise
the person. A system which could correctly detect and recognise human faces would be
useful in a number of important applications including: security and surveillance, robot
vision, and banking. 2D face recognition has been popular to date. However, the large
body of research concentrating on 2D face recognition has not been able to improve the
success in settings such as variation of illumination, variation of pose, makeup, delay
time between gallery and probe. By contrast, 3D techniques may be able to handle
some problems caused by changes in pose, illumination, etc. There exist a number of
methods that are able obtain 3D data and construct a 3D model of a persons face using
computer vision algorithms. The 3D model is then used to recognize the person using an
intelligent recognition algorithm. To be able to achieve real-time recognition, the 3D data
acquisition, the 3D model construction, and the recognition algorithms must be run on a
dedicated hardware platform. Reconfigurable hardware is a system that is capable of
modifying itself while running to meet different application needs. Often, a field-
programmable gate array is used to implement a reconfigurable machine. The aim of
this project is to develop an automated system that will use an array of cameras and a
reconfigurable processing platform to detect human faces and recognise them in real-
time. This project is suitable for a group of four students. The first student will be
responsible for the development of the 3d acquisition module including a camera array,
lighting, and necessary interfacing hardware. The second student will be responsible for
the development of the 3D face model reconstruction algorithm. The third student will be
responsible for the development of detection and recognition algorithms. The fourth
student will be responsible for the development of a reconfigurable architecture which
will host the face detection and recognition algorithms.
Outcome: A portable system that continuously captures images from a camera array,
constructs 3D face models, detects face patterns, and determines the identity of the
person in real-time.
Level of Difficulty: Challenging
Required Background: Students 1&4: Good understanding of digital design techniques
including VHDL, microprocessor and microcontrollers, and low level hardware-related
programming. Student 2&3: Strong programming skills, good understanding of digital
signal processing techniques, and competence in mathematics.
Number of Students: up to 4

KOU-06-3: Intelligent Flower Fragrance Sensor System

Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: Electronic noses can mimic the human sense of smell, and can be used in
a wide range of applications including food processing, environmental monitoring,
medical diagnostics, process control, fragrance development, security, etc. An electronic
nose is generally developed using an array of chemical sensors mimicking the olfactory
receptors, and a data analysis embedded platform which is used to extract quantitative
and qualitative information on the chemical environment. A sensor comprises a material
whose physical properties change according to the concentration of the chemical
material. This change is then translated into an electrical or optical signal. Initially,
different chemicals are presented to the sensor, and a database of signature signals is
built up and labeled. This database of labeled signature signals is then used to train the
data analysis embedded platform. The goal of the training process is to enable the data
analysis embedded platform to produce unique classification of each chemical. The aim
of this project is to develop an electronic nose system that can sense and classify some
flower fragrances. The flower fragrances that will be initially considered include rose,
jasmine, and orange. This project is suitable for a group of two students. The first
student will be responsible for the development of the sensor array, and signal
generation part. The second student will be responsible for the development of the data
analysis algorithm as well as the embedded platform executing the algorithm.
Outcome: A system including hardware and software that senses flower fragrance using
a sensor array, convert it to a suitable signal, classifies the flower fragrance using a
learning system, and also interacts with the user.
Level of Difficulty: Challenging
Required Background: Student 1: Very good understanding of electronics,
measurement devices, and sensors. Students 2: Good understanding of digital design
techniques including VHDL and microprocessor/microcontrollers, good programming
skills, understanding of digital signal processing techniques, and competence in
mathematics.
Number of Students: 2

KOU-06-4: Smart Smell Sensor System for Automatic Classification of Seafood

Freshness
Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: The aim of this project is to develop an electronic nose system that can
sense and classify seafood smell. This project is suitable for a group of two students.
The first student will be responsible for the development of the sensor array, and signal
generation part. The second student will be responsible for the development of the data
analysis algorithm as well as the embedded platform executing the algorithm.
Outcome: A system including hardware and software that senses seafood smell using a
sensor array, convert it to a suitable signal, classifies the sensed smell using a learning
system, and also interacts with the user.
Level of Difficulty: Challenging
Required Background: Student 1: Very good understanding of electronics,
measurement devices, and sensors. Students 2: Good understanding of digital design
techniques including VHDL and microprocessor/microcontrollers, good programming
skills, understanding of digital signal processing techniques, and competence in
mathematics.
Number of Students: 2

KOU-06-5: Miniature Driver Drowsiness Detection System Using Reconfigurable

Hardware
Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: An accident involving a drowsy driver has a high fatality rate due to the fact
that the vehicle control abilities reduce significantly when the driver falling asleep.
Drowsy driver detection systems can reduce the risk of catastrophic accidents related to
drowsy driving. A number of approaches could be used to detect driver drowsiness. The
aim of this project is to: (i) develop an algorithm to analyse eyelid movements and
blinking behaviour of the driver in images captured by a miniature camera and determine
driver drowsiness states ranging from full wakefulness to sleepiness, and (ii) implement
the drowsiness detection algorithm on a reconfigurable architecture. The system
including the camera is placed directly towards the drivers face capturing images. The
implemented algorithm monitors the drivers eyes in order to detect his/her drowsiness in
real-time. The system should be able to provide audio and visual warning to the driver.
This project is suitable for a group of two students. The first student will be responsible
for the development of drowsiness detection algorithm. The second student will be
responsible for the development of the reconfigurable architecture which will interface
with the camera, and also host the drowsiness detection algorithm.
Outcome: A portable miniature system including hardware and software that
continuously captures images from a camera, finds the drivers eyes within images,
analyses eyelid movements and blinking behaviour of the driver, and determines driver
drowsiness states.
Level of Difficulty: Difficult
Required Background: Student 1: Strong programming skills, good understanding of
digital signal processing techniques, and competence in mathematics. Students 2: Good
understanding of digital design techniques including VHDL, microprocessor and
microcontrollers, and low level hardware-related programming.
Number of Students: 2

KOU-06-6: Miniature Intelligent Road-Sign Recognition System Using

Reconfigurable Hardware
Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: In an intelligent vehicle, a driver assistant system monitors the driver,
vehicle, and the road to assist the driver making decisions navigating the vehicle. Road
signs offer information about the traffic conditions. Automated road-sign recognition is
considered an important task in the driver assistance system. A camera captures images
of the road ahead. A computer checks images and intelligently compares them with a
database of road-sign images. The system will have the potential to help in improving
road safety. The aim of this project is to: (i) develop an intelligent algorithm to use a
databases of road-sign images to acquire images from the camera, and detect and
recognise road-signs from within the images in a reasonable time frame, and (ii)
implement the recognition algorithm on a reconfigurable architecture.The system
including the camera is placed directly towards road-signs captures images in real-time.
The system should be able to provide audio and visual warning to the driver. This project
is suitable for a group of two students. The first student will be responsible for the
development of road sign detection and recognition algorithm. The second student will
be responsible for the development of the reconfigurable architecture which will interface
with the camera, and also host the algorithm.
Outcome: A portable miniature system including hardware and software that
continuously captures images from a camera, detects and recognises road-signs within
images.
Level of Difficulty: Difficult
Required Background: Student 1: Strong programming skills, good understanding of
digital signal processing techniques, and competence in mathematics. Students 2: Good
understanding of digital design techniques including VHDL, microprocessor and
microcontrollers, and low level hardware-related programming.
Number of Students: 2

KOU-06-7: Classification of Attractiveness in Facial Images

Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: Attractiveness is an emotional factor that can have an effect on how
people are judged, in terms of employment or social opportunities, friendship, etc. The
aim of this project is to develop an intelligent algorithm to learn the attractiveness
preferences of a person using a set of a few example face images, and then
automatically classify a large collection of face images using the learnt attractiveness
preferences. The algorithm will be implemented in C, C++, or Matlab. The program will
include a GUI, training and test face databases, and learning and classification
components.
Outcome: Software that presents a few prespecified face images to a person, learns
attractiveness preferences of the person, then classifies and presents to the person the
face images of the attractive people from within a large collection of face images.
Level of Difficulty: Difficult
Required Background: Strong programming skills especially in MATLAB, C or C++,
good understanding of digital signal processing techniques, and competence in
mathematics
Number of Students: 1

KOU-06-8: Voice Morphing

Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: Voice morphing is a method which is used to modify a source speaker's
speech characteristics to sound as if it was spoken by a target speaker giving it a new
identity while preserving its original content. The aim of this project is to develop an
algorithm which can morph speech from a source speaker to a target speaker. For
example, the algorithm could modify the spectral properties and pitch of the source voice
to match those of the target voice. The algorithm will be implemented in C, C++, or
Matlab. It will capture live voice data from a microphone which is connected to a PC,
transform the voice, and play back the transformed voice on a speaker attached to the
PC.
Outcome: Software that captures live voice data, transforms the voice data, and plays
back the transformed voice on a speaker.
Level of Difficulty: Difficult
Required Background: Strong programming skills especially in MATLAB, C or C++,
good understanding of digital signal processing techniques, and competence in
mathematics
Number of Students: 1

KOU-06-9: Automatic Detection of Deception

Proposed By: Dr Abbas Kouzani
Industry Partner: None
Description: People tell lies for different reasons. Classical lie detectors are instruments
that test three aspects of the human physiology: respiration, changes in skin
resistance/condeuctance, and blood volume and pulse rate. Today, portable non-contact
lie detectors are being developed that use voice recognition technology to determine
stress, tension and excitement in a person's voice. Also, it has been observed that EEG
signals may be useful in identifying deceptive behavior. The aim of this project is to
develop a lie detector machine to utilise such parameters as physiological, voice, EEG
data to automatically specify whether deceptive behavior is being displayed or not. The
system will collect the physiological, voice, and EEG data, measure variations in the
collected data, and specify automatically whether deceptive behavior is being displayed
or not.
Outcome: Hardware and software components of a lie detection system that is capable
of specifying automatically whether deceptive behavior is being displayed or not.
Level of Difficulty: Difficult
Required Background: Good understanding of electronics, measurement devices,
sensors, and microprocessor/microcontrollers, good programming skills, understanding
of digital signal processing techniques, and competence in mathematics.
Number of Students: 1

Brendan McGrath

McG-06-1: Design and Modelling of a 5kW Zero Voltage Switched Full Bridge
Converter for Railway Applications.
Proposed By : Brendan McGrath.
Industry Partner: Not Applicable.
Description: Modern diesel powered locomotives often make use of high current
capacity battery banks as a back-up power source when the diesel alternator is not
available (e.g. start-up conditions). These locomotives therefore require electrically
isolated, high power throughput battery chargers. A key challenge in traction
applications, however, is to reduce the size and weight of the required auxiliary
systems while simultaneously operating in difficult environments (i.e. high ambient
temperatures, 20% voltage fluctuation, high vibration, etc.). One way to reduce the
size of the battery charger is to use a power electronic solution at a high switching
frequency which reduces the size of the magnetic components.
One candidate converter for this type of application is the Zero Voltage Switched (ZVS)
Full Bridge Converter. This structure uses a H-Bridge with capacitors across each switch
to drive a transformer, and these capacitors form a resonant circuit with the transformer
leakage inductance allowing for soft-switching. The transformer output is then rectified
using either a centre-tapped rectifier arrangement, or a current doubler circuit. This
project is a study into the feasibility of using the ZVS converter in this type of application.
It will be necessary to design a converter at representative voltage and power levels,
and this includes magnetic, thermal and power electronic designs. It will then be
necessary to model the converter using a simulation tool such as Saber, and if possible
investigate secondary effects on the converter operation such as tail-current losses in
the devices. A significant component of this project will involve a literature survey.
Outcome : A complete design of the ZVS converter, and simulation models that can be
used to investigate the performance of the converter in a wide range of operating
conditions.
Level of Difficulty : Difficult.
Required Background : This project requires a background in power electronics, and
electro-magnetics.
Number of Students : This project is designed for one student.

McG-06-2: Design and Modelling of a Multilevel Shunt Active Filter for Medium
Voltage Distribution Networks.
Proposed By : Brendan McGrath.
Industry Partner: Not Applicable.
Description: Non-linear loads connected to the AC grid are problematic because they
inject harmonic currents into the grid. Common example loads include switched
mode power supplies, variable frequency drives, and saturating magnetic equipment
(i.e. transformers). This distortion can lead to a range of unwanted side-effects,
including increased system losses, increased dielectric stress, and malfunction of
susceptible equipment. Active harmonic filters have been proposed as one solution
to the problem. These devices are placed in parallel (i.e. shunt) with non-linear
loads, and use active cancellation of the harmonic currents drawn by the load,
thereby harmonically isolating the non-linear load from the rest of the grid. While this
approach has been demonstrated for low voltage distribution networks, there remain
many questions concerning the efficacy of the approach for medium voltage (i.e.
6.6kV and above) networks. In particular the higher voltages involved require
alternative converter topologies compared to the conventional two-level approach,
and multilevel (i.e. Cascaded, Diode Clamped and Flying Capacitor structures) have
shown some promise in this area.
The aim of this project is to design a medium voltage multilevel converter suitable for an
active filter application, and then model the converter using an appropriate simulation
tool (Saber) under a range of operating conditions. Some of the considerations to be
investigated include the required converter rating, the number of harmonics that can
reasonably be eliminated, and any additional components that may be required to
ensure satisfactory performance of the converter. There will be a significant literature
study required for this project.
Outcome : A design for a multilevel inverter that can be used as a shunt active filter at
medium voltage levels. Complete simulation models, that can be used to investigate
the performance of the active filter in a wide range of operating conditions.
Level of Difficulty : Difficult to Challenging.
Required Background : This project requires a background in power electronics and
power systems.
Number of Students : This project is designed for one to two students.

McG-06-3: Development of a 600V Boost Power Factor Correction Converter.

Proposed By : Brendan McGrath.
Industry Partner: Not Applicable.
Description: It is well known that grid connected power supplies which utilise single
phase diode rectifier circuits inject harmonic distortion into the AC grid. This
distortion is problematic, and can lead to poor power factor, voltage distortion,
conductor heating, and malfunction of electrical equipment connected to the grid.
Power Factor Correction (PFC) rectifiers offer one solution to the problem, and
make use a boost converter coupled to the output of the single phase rectifier. When
the boost circuit is switched at high frequency (e.g. 10kHz), the resulting rectifier
input current contains only high frequency switching ripple, and the power factor is
close to unity.
PFC circuits also offer the potential to boost the rectified DC bus voltage to well above
the theoretical limit of 340V achievable by single phase rectifiers, and this makes PFC
circuits suitable to a wide range of additional applications. For example at a DC bus of
600V it is possible to couple the PFC rectifier to a three phase 415V AC drive, and so
the PFC circuit operates as a high performance single phase to three phase converter.
However it is challenging to build PFC circuits operating at 600V, because the required
high voltage boost diodes have significant reverse recovery limitations that limit the
switching frequency. One way around this limitation is to use a three level boost
topology, but the control of this multilevel structure is far more complex. In this project it
is required to examine the two competing PFC solutions, and then a 600V PFC rectifier
using one of these structures will be constructed. There will be a significant literature
study required for this project.
Outcome : A complete design of the PFC circuit, simulations of the system, and a
working prototype PFC (two or three level), that can be run up to 600V DC at up to
500W of load.
Level of Difficulty : Difficult to Challenging.
Required Background : This project requires a background in power electronics and
electro-magnetics.
Number of Students : This project is designed for one to two students.

McG-06-4: Cell Voltage Regulation for Flying Capacitor Converters.

Proposed By : Brendan McGrath.
Industry Partner: Not Applicable.
Description: Flying capacitor multilevel converters are showing considerable promise in
high power applications, such as drives and grid connected converters, largely
because of the ability to operate at high input voltages, while producing low levels of
switching distortion. However for these converters to be viable, a control strategy
must be employed to regulate the voltage across each floating capacitor. While this
can be achieved in open loop with a suitable switching strategy, the dynamics tend
to be quite slow and hence a closed loop method is attractive.
This project involves the investigation of closed loop regulation methods for flying
capacitor converters. A literature review will be conducted to identify candidate
techniques and modelling approaches. Simulations will be developed to investigate each
technique, and a low voltage MOSFET based prototype will be built to experimentally
test the most promising method(s).
Outcome : Simulation models of the flying capacitor converter under a range of
conditions, and a prototype low voltage flying capacitor converter.
Level of Difficulty : Difficult to Challenging.
Required Background : This project requires a background in power electronics and
control systems.
Number of Students : This project is designed for one students.

Brett Ninness

NIN-06-1: Model Predictive Control of an Inverted Pendulum

Proposed By: Brett Ninness
Description: An inverted pendulum is a standard but difficult control problem. This
project involves controlling one in the Signal Processing Microelectronics Laboratory
using a technique called model predictive control (MPC). This will involve learning about
advanced methods of Automatic control and will require programming in C to implement
the controller on a Texas-Instruments platform.
Level of Difficulty: Challenging
Number of Students: 1

NIN-06-2: Model Predictive Control of a Crane

Proposed By: Brett Ninness
Description: The Crane Control Problem involved moving a load as quickly and
precisely as possible while avoiding excess swaying of the load. This project involves
controlling a model crane in the Signal Processing Microelectronics Laboratory using a
technique called model predictive control (MPC). This will involve learning about
advanced methods of Automatic control and will require programming in C to implement
the controller on a Texas-Instruments platform.
Level of Difficulty: Challenging
Number of Students: 1

NIN-06-3: Altera PLD implementation of Multi-user detector

Proposed By: Brett Ninness
Description: Some recent research by myself, Dr. Steve Weller and a Ph.D student
Soren Henriksen has revealed new methods for reception of CDMA modulated signals
that, in simulation, appear to out-perform any current competing methods. We want to
progress to testing these new methods in practice, and also to establish their feasibility
for actual chip implementation. This project will contribute to that research by
implementing the new reception (formally, detection) methods on a PLD. This project is
challenging. It will suit someone who wants to progress to a career in hardware design,
particularly with a focus towards telecommunications.
A successful implementation of this project would feed into a long-term project within the
Signal Processing Microelectronics (SPM) research group with the School of EE&CS to
construct a wireless communications testbed.
Level of Difficulty: Challenging
Number of Students: 1

NIN-06-4: Wireless Testbed Driver and GUI-based Interface

Proposed By: Brett Ninness
Description: The Signal Processing Microelectronics (SPM) research group has
developed a wireless testbed. This project involves writing a Windows driver and
associated interface to control the testbed. This will require programming in C to
implement the driver, and in the students language of choice to implement the GUI.
Level of Difficulty:Moderate to Challenging
Number of Students: 1

NIN-06-5: Web Control/Monitor

Proposed By: Brett Ninness
Description: This project involves designing and building a unit which can monitor
various digital and analog inputs, and report their values via a web based interface.
Additionally, this same web based interface should be able to control digital and analog
outputs on the device. A typical application for such a device would be a home
automation system which can be controlled and monitored from a web browser.
Level of Difficulty: Easy/Moderate
Number of Students: 1
NIN-06-6: GPS Controlled Clock
Proposed By: Brett Ninness
Description: This project involves building a highly accurate clock, which uses GPS
signals for timebase and calibration. This involves receiving and decoding a GPS signal,
and then extracting appropriate time information for further processing by a
microcontroller to implement the clock.
Level of Difficulty: Easy
Number of Students: 1

NIN-06-7: Dive Computer

Proposed By: Brett Ninness
Description: Scuba diving requires computation of allowed times at given depths in
order to avoid injury due to excess nitrogen being dissolved in the blood - ie, to avoid
"the bends". This can be computed before a dive, using a table (this is sorta
compicated), or by using an expensive submersible "dive computer". The purpose of this
project is to develop a surface-used device which is a "dive planner" and logger. The aim
would be to build the cheapest possible piece of consumer electronics that would allow a
recreational diver to easily develop a plan of multiple dives such that no decompression
limits were breached. This is a relatively easy project, but could move towards the
moderate level if a very comprehensive and complete solution were developed. It would
suit someone interested in developing embedded systems with a mixture of hardware
and software design.
Level of Difficulty: Easy
Number of Students: 1

NIN-06-8: MultiChannel Radio Control Unit

Proposed By: Brett Ninness
Description: This project involve designing and building a radio control transmitter and
receiver pair that can be used for control of multiple devices.
Level of Difficulty: Easy
Number of Students: 1

Peter Schreier
SCH-06-1: Web experiments in digital communications
Proposed by: Peter Schreier
Description/Outcomes: This project is about developing some interactive
demonstrations/experiments/visualizations in digital communications accessible on the
Internet. As such, it requires some familiarity with digital communications, but the project
itself is more of a programming project than a communications project. To give you an
idea of how these experiments could work, take a look at
http://www.jhu.edu/~signals/index.html from Johns Hopkins University.
Most of the experiments on this webpage are written in Java, and that seems to be a
reasonable choice to me (simple HTML will not be good enough). Thus, at the end of this
project, you should have a webpage with a few experiments/demonstrations in digital
communications. Some ideas are: eye diagrams, power spectra of digital modulation
techniques, illustration of the binary symmetric channel, but you are definitely not limited
to these. The more creative you are and the more sophisticated your experiments are,
the higher your final grade will be.
Skills Required/Developed: This project requires a good background in digital
communications and experience with Web programming techniques (probably Java).
ELEC3530 /3520
Level of Difficulty: Moderate
Number of students: 1-2

SCH-06-2: Time-frequency toolbox for MATLAB

Proposed by: Peter Schreier
Description/Outcomes: In this project, you will develop a MATLAB toolbox for time-
frequency analysis of nonstationary signals. This toolbox should contain
implementations of several time-frequency representations. These will be tested on
some practical data from real-life applications.
Skills Required/Developed: This project requires a very good background in
mathematics (linear algebra). ELEC2400
Level of Difficulty: Difficult
Number of students: 1-2

SCH-06-3: Performance evaluations of CMDA systems

Proposed by: Peter Schreier
Description/Outcomes: In this project, you will implement a number of different
receivers for a CDMA system (single-user, decorrelator, MMSE, decision feedback, )
in MATLAB and compare them in terms of their performance (through simulation). This
will be done for AWGN channel and fading channels.
Skills Required/Developed: This project requires a very good background in
mathematics (linear algebra) and digital communications. As a rule of thumb: If you dont
know what the inverse of a matrix is or how QPSK works, dont choose this project.
ELEC3530
Level of Difficulty: Difficult
Number of students: 1-2

SCH-06-4: Vehicle tracking with linear dynamic models

Proposed by: Peter Schreier
Description/Outcomes: Systems that can track cars using video from fixed cameras
can be used to predict traffic volume and flow. The ideal is to report on and act to
prevent traffic problems as quickly as possible. In this project, you will implement a
number of tracking systems based on linear dynamic models (such as Kalman filtering
and data association), and evaluate and compare their performance.
In this project, you will gain some exposure to modern computer vision problems and
solutions.
Skills Required/Developed: Familiarity with probability, linear systems, and linear
algebra is required. ELEC2400 ELEC3400
Level of Difficulty: Moderate to difficult
Number of students: 1
 Terry Summers

SUM-06-1: Single-Phase Adjustable Speed Motor Drive

Proposed By : Terry Summers
Description: This project is derived from the Future Energy Challenge
(http://www.energychallenge.org) of 2004. The problem as defined by the challenge is,
Innovations in motors and motor drive systems that produce deep cuts in losses and
costs for home (appliance) use, or that could replace universal motor brush machines
in residential applications. Target hardware costs are US$40 for a combination of motor,
power electronic driver, and controller that can operate from a single-phase residential
source, deliver rated shaft load of 3/4 HP (or 500 W) at 1500 RPM, exhibit a useful
speed control range of at least 150 RPM to 5000 RPM, and provide power efficiency of
at least 70% for loads ranging from 50 W to 500 W at a specified speed.
This project was offered in 2005 with the outcome being a reasonably sophisticated
design and proof of concept using a permanent magnet synchronous machine as the
motor. The aims of the project in 2006 are to:
Complete the Saber simulation using, where possible, models of real parts so that the
overall efficiency and of
the performance of the system can be accurately estimated.
Design build and test components of the drive system including, but not limited to, the
front end power factor
correction boost converter and three phase inverter.
Students will not be limited to considering only the existing design and in fact are
encouraged to suggest improvements
to the existing solution or a completely new solution of their own.
Outcomes: A working single phase adjustable speed drive.
Level of Difficulty : DifficultChallenging
Number of Students : 1

SUM-06-2: Utility Interactive Inverter System for Small Distributed Generation

Proposed By : Terry Summers
Description: This project is derived from the Future Energy Challenge
(http://www.energychallenge.org) of 2004. The project is being run as a trial with view to
potential participation in the competition in future years. The problem as defined by
the challenge is, The objective of this competition is to improve and foster innovation in
the design of flexible utility interactive inverter systems for small distributed generation.
The inverter must be efficient and comply with requirements for harmonic control,
interconnection, and safety standards. The inverter must be capable to operate in typical
voltage and frequencies, grid-connected or stand-alone. Target hardware costs are
US$200 for an inverter that can operate from a dc input voltage varying from 30 volts to
60 volts, under constrained input power varying from 250W to 1000W, to a single-phase
utility line 110/240V either 50Hz/60Hz, with efficiency of at least 90%.
Outcomes : A working single phase inverter.
Level of Difficulty : DifficultChallenging
Numbers of Student: 1
SUM-06-3: Battery Condition Monitoring System
Proposed By : Terry Summers
Description: Battery condition monitoring and smart charging systems are often
important elements in large Uninterruptable Power Supplies (UPSs) due to their ability to
increase battery life, reduce maintenance costs and improve system reliability.
Monitoring systems may well become an important part of automotive and small
distributed renewable energy systems for the same reasons.
This project aims to design and develop a battery monitoring system for the battery bank
which forms part of the dc supply to the the electrical machines laboratory. At a minimum
the monitoring system should be able to provide information regarding the condition of
individual batteries via an automated test procedure. Possible extensions of the
project could involve the design of an improved charging system for the batteries to
improve their lifetime and hence reduce their lifetime cost.
Outcomes : A working battery monitoring system.
Level of Difficulty : ModerateChallenging
Numbers of Student:1

SUM-06-4: Power Supply

Proposed By : Terry Summers
Description: There is a demonstrated need for a power amplifier in the Electrical
Engineering Laboratory which is able to operate at reasonable voltage levels and
currents as well as have a relatively wide bandwidth.
This project requires a student to design an build a power amplifier which can operate at
fundamental frequencies ranging from dc to 600Hz, voltages up to 240 Vrms and
currents up to 8 Arms.
Outcomes : A working power supply.
Level of Difficulty : ModerateChallenging
Number of Students : 1

SUM-06-5: Programmable Ignition System for LPG and/or Dual Fuel Cars
Proposed By : Terry Summers
Description: Dual fuel (petrol / liquid petroleum gas (lpg)), conversions of older, less fuel
efficient automobiles, to take advantage of both the lower price and the environmental
benefits of lpg, are fairly popular in Australia and overseas. Unfortunately
most conversions dont take into consideration the different ignition requirements of lpg
and consequently dont take full advantage of the fuel. LPG conversions then tend to
achieve far less than optimal performance and fuel economy from the available motor.
A number of after market ignition systems exist in the marketplace that cater for dual
ignition curves however, most are either too expensive, have limited performance or
require substantial modification to the existing ignition system.
This project requires the design and manufacture of an dual fuel ignition system which
will allow a vehicle to operate on its standard timing map whilst operating on petrol and
a timing map optimised for lpg when operating on lpg. This system should ideally be able
to be easily retrofitted to cars which have had dual fuel or lpg only conversions carried
out.
Outcomes : An operating ignition system for passenger cars.
Level of Difficulty : Easy - Challenging
Number of students : Suitable for one student
SUM-06-6: Control of a Electromagnetic Vibratory Feeder
Proposed By : Terry Summers
Industry Sponsor : Maxon Machinery (Stuart Adams)
Description: Electromagnetic vibratory feeders are widely used in manufacturing
industries for the transportation of solid materials.
Feeders are quite different from conveyors. Conveyors are only capable of transporting
goods while vibratory feeders are used to control the rate at which materials flow. Typical
applications include feeding to packaging, batching, mixing, drying and similar
applications. Current tests on the feeder show that the feeder suffers from major power
factor inefficiencies and other electromagnetic problems.
The purpose of this project is to modify/redesign the electronics on a electromagnetic
vibratory feeder provided by Maxon Machinery to alleviate these problems.
Outcomes :Working electromagnetic vibratory feeder where:
The feeder operates at or close to unity power factor.
Efficient control of vibrational feeder to control feed rate is achieved.
Level of Difficulty : Moderate - Challenging
Number of Students : Suitable for one student

Steve Weller

WEL-06-1: An FPGA-based fading wireless channel emulator

Proposed By: Steve Weller
Industry Partner: none
Description: The aim of this project is to implement a device based around a field-
programmable gate array (FPGA) for emulating the behaviour of a wireless
communication channel. Due to the effects of constructive and destructive interference
in multipath wireless channels, the received signal power is subject to fluctuations known
as fading. A very common model for this fading is to assume so-called Rayleigh fading,
in which the real and imaginary components of a complex baseband signal model are
described by Gaussian random variables.
The basis for this project is therefore to construct a digital noise generator whose
probability density function (PDF) closely approximates that of a Gaussian random
variable. A description of such a noise generator is provided in the following paper:
D.-U Lee, W. Luk, J.D. Villasenor, and P.Y.K. Cheung, "A Gaussian noise generator for
hardware-based simulations", IEEE Transactions on Computers, vol.53(12), pp.1523-
1534, December 2004. This paper gives a detailed description of how to implement a
Gaussian noise generator using field-programmable gate arrays FPGAs, capable of
speedups of up to 50-times over 2.6GHz Pentium-based implementations, using
comparatively slow and inexpensive FPGAs.
A successful implementation of this project would feed into a long-term project within the
Signal Processing Microelectronics (SPM) research group with the School of EE&CS to
construct a wireless communications testbed. You can read more about the testbed
project and the SPM research group at: http://sigpromu.org
Skills required/developed: programming FPGAs with graphical entry and/or VHDL
Outcomes: functioning prototype.
Level of Difficulty: moderate/difficult
WEL-06-2: Trellis-based block decoding on a DSP chip
Proposed By: Steve Weller
Description: Forward error correction (FEC) allows the protection of digital data from
channel-induced errors. In this project, data will be encoded and decoded using linear
block codes represented by trellis diagrams. The aim of this project is to implement
trellis-based decoders on a digital signal processing (DSP) chip. The starting point for
this project will be the Viterbi algorithm. A more ambitious project solution will consider
the so-called "BCJR" (or forward-backward) algorithm. In addition, a simple digital noise
source will be simulated, allowing the benefits of forward error correction to be clearly
demonstrated.
A successful implementation of this project would feed into a long-term project within the
Signal Processing Microelectronics (SPM) research group with the School of EE&CS to
construct a wireless communications testbed. You can read more about the testbed
project and the SPM research group at: http://sigpromu.org
Skills required/developed: programming in C; understanding of simple forward error
correction techniques and trellis-based representations of codes
Outcomes: forward error correction encoder and decoders implemented
on a suitable DSP development platform
Level of difficulty: moderate/difficult

WEL-06-3: Implementation of cyclic redundancy check (CRC) codes on an FPGA

Proposed By: Steve Weller
Description: Cyclic redundancy codes (CRCs) are a simple, widely used and extremely
effective means of controlling errors in digital communication systems. Rather than
attempting to correct errors, CRC codes simply append check bits to enable errors to be
detected at the receiver. Blacks which are received in error can then be re-transmitted.
The goal of this project is to use the CRC-8 algorithm to enable a sender to calculate
and append a CRC byte to the end of a data packet. When the receiver gets the data, it
too calculates its own CRC byte and then compares it to the incoming CRC byte. If the
two CRC bytes do not match, an error flag is raised. The Altera UP1 (or UP2) board will
be used as the implementation platform.
A successful implementation of this project would feed into a long-term project within the
Signal Processing Microelectronics (SPM) research group with the School of EE&CS to
construct a wireless communications testbed. You can read more about the testbed
project and the SPM research group at: http://sigpromu.org
Outcomes: forward error correction encoder and decoders implemented on an Altera
CPLD/FPGA development board
Skills required/developed: programming FPGAs with graphical entry and/or VHDL;
understanding of simple forward error correction techniques
Level of difficulty: moderate

WEL-06-4: A codec for the binary erasure channel

Proposed By: Steve Weller
Industry Partner: none
Description: In many communication channels, bits are received which are noise-
corrupted. Sometimes 0s are received as 1s (or vice versa). Sometimes the received
symbols are corrupted with Gaussian noise. In binary erasure channels, however, bits
are either received perfectly or are totally wiped out (erased) with some probability. This
sounds like a strange state of affairs, but it is a very good model for Internet
communications, where packets mostly arrive error-free, but occasionally don't arrive at
all.
In this project, the aim is to produce an encoder/decoder (codec) for the binary erasure
channel, using a class of error-correction codes known as low-density parity-check
(LDPC) codes.
There are two possible implementation platforms for this project: either field-
programmable gate arrays (FPGAs), or on a digital signal processing (DSP) platform.
Either way, I'm interested in a solution which enables the erasure correction process to
be readily visualised. Using the Altera UP1/UP2 FPGA development system for example,
it is possible to generate a VGA video display signal directly from the UP1/UP2.
A successful implementation of this project would feed into a long-term project within the
Signal Processing Microelectronics (SPM) research group with the School of EE&CS.
You can read more about the testbed project and the SPM research group at:
http://sigpromu.org
Outcomes: erasure-correcting encoder/decoder implemented on an Altera CPLD/FPGA
development board or suitable DSP development system
Skills required/developed: programming FPGAs with graphical entry and/or VHDL;
understanding of simple forward error correction techniques; C programming
Level of difficulty: moderate/difficult

WEL-06-5: Hardware implementation of a Reed-Solomon codec.

Proposed By: Steve Weller
Industry Partner: Agere Systems (industry-based supervisor: Dr. Adriel Kind)
Description: The physical environments in which point-to-point communication typically
takes place are prone to introduce corruption into the transmitted data. Examples of
such impairments are light dispersion in an optic-fiber, multipath echos in urban wireless
environments, or random bit-flipping due to ionisation on a magnetic disk.
Reed-Solomon codes are interesting and very powerful mathematical constructions,
which provide theoretically optimal protection against errors. The codes are traditionally
used in practice for correcting packet errors in networking and data storage applications.
Decoding of the packet at the receiver is usually performed using the Berlekamp-Massey
algorithm, which can be quite computationally complex.
The aim of the project is to investigate and propose a digital decoding circuit, which may
be implemented in a practical receiver.
Outcomes: An architecture proposal for a digital Reed-Solomon encoder/decoder, along
with functional Matlab code demonstrating a working solution. A very capable student
taking this project will also have the opportunity to implement a real circuit layout using
Cadence hardware description/design tools.
Skills required/developed:
Understanding of error correction coding and decoding.
Computational complexity issues.
Implementation of complicated algorithms in software.
Possibility of digital hardware development.
Level of difficulty: moderate/difficult
 James Welsh

WELSH-06-1: Low Cost Personal Monitoring System

Description: The idea of this project is to design and construct a low cost personal
monitoring system. The system should be designed around a microprocessor to which
specific measuring devices are connected via a suitable network topology.
Measurements to be considered include mass, height, temperature, heart rate, blood
pressure, oxygen saturation, lean body mass, etc. The system should automatically
detect when measuring devices are connected and disconnected from the network.
Measurements should be displayed in an appropriate form on a small LCD.
Outcome: This project involves the design of a microprocessor system and the
electronics associated with the sensors to conduct the various
measurements. Software development will include code to read data
from the sensors, process this data and display the readings.
Difficulty: Easy to challenging.
Number of students: 1

WELSH-06-2: Microprocessor Controlled Synchronous Machine

Description: In this project the student is required to design and implement a
microprocessor based system for the control of a synchronous machine. The function of
the controller will be to automatically connect the machine to the mains power supply
and, once connected, provide power factor correction at a desired level for a local load.
This project will utilise the school's 42 Volt laboratory scale machines.
Skills / Background: Skills in the broad areas of microprocessor systems, electronics
and AC machines will be required.
Difficulty: Difficult.
Number of students: 1

WELSH-06-3: Electric Nail Gun

Description: This project involves research into the feasibility of an electric nail gun as
to whether it can perform, in theory, as well as a pneumatic nail gun (typical measures of
performance would include size, mass and energy requirements). The student will also
be required to design and construct a prototype electric nail gun to verify their research.
Skills / Background: This project will require skills in both electromagnetic and
electronic design.
Difficulty: Easy to difficult.
Number of students: 1

WELSH-06-4: Magnetic Levitation Train

Description: This project involves the design and construction of a scale model
magnetic levitation train. A magnetic leviation train presents a number of challenges in
hardware design, sensing and control. In this first year of offering it is envisaged that
majority of the student's effort will be towards the infrastructure and propulsion
system for the train.
Skills / Background: This project will require skills in electromagnetic and electronic
design as well as a knowledge of microprocessor systems.
Difficulty: Difficult to challenging.
WELSH-06-5: Magnetic Levitation Demonstrator
Description: Magnetic levitation demonstrators are typically used in laboratory
exercises for demonstrating the theory learnt in control courses. A common
demonstrator will levitate an object at a set distance above or below an electromagnet.
The goal of this project is to design and construct a magnetic levitation demonstrator that
will serve two purposes: 1). to be utilised in ELEC4410 Control System Design and
Management Laboratories and 2). to be used as a promotional tool for the School of
Electrical Engineering and Computer Science. It is a requirement of the project that the
magnetic levitation device be a complete stand alone unit.
Skills and Background: This project will require skills in electromagnetic and electronic
design as well as a knowledge of microprocessor systems.

WELSH-06-6: Electronic Nature Guide

Proposed by: Steve Mitchell.
Description: The basis of this project is to design and construct an electronic nature
guide that will ``listen'' to the sounds of the bush and identify calling animals. Once
identification has been made the results are to be displayed on a colour, graphical LCD
showing an
image and detailed information on the animal species.
The device should be a small battery powered module including a processor capable of
doing the necessary signal processing for ``call signature'' identification. Other features
could include recording call which were not identified, for future expansion of
the call signature database.
This project will require skills in digital signal processing and microprocessor systems.
Difficulty: Moderate to difficult.

WELSH-06-7: Ball and Plate

Description: This project initially involves revisiting the mechanical and electrical design
of the existing ball and plate system to address some inadequacies. The ultimate goal of
the project is the implementation of a control system that enables the position of the ball
to be maintained at a desired location on the plate.
Level of Difficulty: Challenging.

WELSH-06-8: Multivariate statistical modelling methods for prediction of mineral

processing variables from surface vibration monitoring data
Industry Partner: CSIRO
Description/Outcomes: A large variety of devices are utilised in the processing of raw
ores for the production of metals and preparation of coal for power station feed. These
include comminution (breakage) and classification (separation) unit operations used to
concentrate the components of interest. For instance, Dense Medium Cyclones (DMCs)
are commonly used devices in coal preparation plants for separation of pure coal from
gangue (waste) rock in slurries.
CSIRO Minerals is investigating the use of surface vibration (contact Acoustic Emission -
AE) monitoring in conjunction with advanced signal processing and soft-sensor
inferential data analysis and modelling techniques based on multivariate statistical
analysis to develop non-intrusive process and equipment condition monitoring systems
for mineral processing devices.
A number of data-driven multivariate statistical analysis methods are being investigated
for analysing features extracted from multi-channel contact AE measurements and
formulating quantitative models for process and condition variables of interest. These
methods include Principal Component Analysis (PCA) / Principal Component Regression
(PCR) and Partial Least-Squares / Projection to Latent Structure (PLS). The signal
features input to the data-driven modelling methods include time series statistical
characteristics and Fourier (frequency domain) parameters. Wavelet (time and scale)
data analysis has also been investigated as an effective feature extraction technique for
mineral processing equipment contact AE containing sharp (transient) events. The
technique of wavelet analysis has been successfully applied to mineral processing
equipment contact AE measurements in an Honours year project (Woodham, 2004) and
a vacation scholarship project (Dave, 2004). Wavelet analysis combined with Fisher
Discriminant Analysis (FDA) has been successfully used to construct qualitative
(classification) models for process state in vacation scholarship and Honours year
projects (McElroy, 2005a,b).
The Honours project will consist of two components, with the first component being the
base level and the latter a project extension.
Base Project: Development of quantitative multivariate statistical modelling software
that utilises wavelet, Fourier and time series features extracted from mineral processing
equipment AE data
Adapt existing PCR and PLS software to develop quantitative models for mineral
processing state based on selected features extracted from contact AE
measurements by wavelet, Fourier and time series analysis. The software will be
based on the MATLAB technical programming development environment.
Develop a regression-based quantitative predictive modelling method based on FDA.
Investigate the use of a variety of AE feature variables as inputs to the method.
Preparation of a report.
Extension: Investigation of the relative accuracy of a variety of multivariate statistical
modelling methods for quantitative prediction of mineral processing state from AE data
sets
Apply the multivariate statistical analysis quantitative predictive modelling software
developed in the base project to a selection of mineral processing equipment AE
data sets for a wide variety of operating conditions. Assess the relative accuracy of
the methods for prediction of the variables of interest.
Establish whether any specific combination of feature variables results in more
accurate predictive models for process variables, using PCR, PLS and FDA based
multivariate statistical analysis methods.
Level of Difficulty: Challenging
Number of Students: 1

NOTE: It is expected that the student will provide short progress reports to CSIRO
Minerals once per month in the first session of the university year and will have direct
contact (visits to the Lucas Heights site) on average at least once every two weeks
during the second session. Travel expenses will be paid by CSIRO

Mehmet Yuce
YUC-06-1: Analysis and Implementation of High Speed Costas Loop Type PLLs
Description: Phase-locked loops (PLLs) have been widely used in wireless
communications and many other electronics systems for detection and synchronization.
They have been mostly modeled in S-domain by considering the inputs as analog. In
recent data communication systems, high speed analog-to-digital converts (ADCs) can
be employed to digitize input analog signals. Therefore a PLL can be designed via high
speed digital signal processing. In addition, most of currently available PLL implements a
digital phase-frequency detector (PFD) at the front-end of the loop. These PLLs are
being designed as partially or all-digital because a digital design is more flexible and can
be on-chip with other communication circuits. If a PLL includes some digital
components, the phase information is contained in the data transitions. And they should
be more accurately modeled as discrete-time sequence. This project requires the
stability analysis and implementation of high-speed Costas Loops in discrete-domain
using either Ztransform or discrete-time Fourier transform (DTFT) and compare them
with the most common continues-time model (S-domain). A complete set of equations
needs to be derived for optimum phase acquisition. After verifying the operation of the
system in Spice or Simulink, it should then be implemented in hardware to assure the
validity of the model.
Skills required/developed: Background in Communication or Control Theory or DSP is
required. ELEC3530 ELEC3400
Level of Difficulty: Moderate to difficult
Number of Students: 1

YUC-06-2: Coil Optimization Techniques for Power and Data Transfer in Medical
Implantable Devices
Proposed by: Mehmet R. Yuce
Description: Inductive links via two coil pairs are commonly used to power electronics
devices in biomedical and RFID applications. The optimization of the link is very crucial
due to the requirement of low-power consumption. Inductive link consists of a primary
coil which is outside the body and a secondary coil implanted in the body. There is a
wide range of transmission bands (i.e. ISM bands) available for medical applications.
The design of coils for each frequency band requires different geometry and optimization
techniques. The student/s need to develop a technique to optimize an inductive link and
minimize the size of coils for ISM bands. Meanwhile it is desired to maintain the
efficiency and the bandwidth requirements. Vision prosthesis can be taken as the
application for this design. First a software program should be developed to include
different shapes and materials for optimum coil design. The technique should also be
experimentally tested and evaluated in the lab.
Skills required/developed: some electromagnetic theory as well as some
communication theory, and electronics. ELEC3240 or ELEC3210, PHY3500
Level of Difficulty: Moderate -Difficult
Number of Students: 1

YUC-06-3: An Intelligent Bi-directional Wireless Telemetry Design for Vision

Prosthesis
Proposed by: Mehmet Yuce with Gregg Suanning (Mechatronics)
Description: Retinal prosthesis an implantable prosthetic system is under
development at several universities around the world to restore vision in people who
have profound loss of vision due to degeneration of the photoreceptors in the retina
caused by diseases such as RP (Retinitis Pigmentosa) and AMD (Age related Macular
Degeneration). The prosthesis consists of an external microelectronics system that
captures and processes image data from a video camera outside the body and transmit
the information as electrical signal to an integrated circuit embedded inside the body
(typically eye). The goal of this project is to design a radio frequency (RF) power and
data link to provide communication between the camera and the implanted electronics
and to transfer power wirelessly through inductively coupled coil pairs. The project will
develop several new techniques that might be applicable to a vision telemetry system. It
should also have the capability to monitor activation/stimulation on the implant site from
the outside.
Skills required/developed: Electronics and wireless communication. ELEC3240 or
ELEC3210
Level of Difficulty: Difficult
Number of Students: 1 or 2

YUC-06-4: Wireless Infrared Communication for Retinal Prosthetic Device

Proposed by: Mehmet Yuce with Gregg Suanning (Mechatronics)
Description: A wireless optical transmission system is to be developed for Retina
Prosthesis. Unlike the common RF link, here the image data from the video camera will
be transmitted via an optic link to the implanted device in the body. The technique will
include an infrared LED with a suitable wavelength as a transmitter and a photodiode
(PD) to receive the data and to reconvert it to an electrical signal. The data demodulation
part takes the output of PD into digital signal for further signal processing. An appropriate
modulation scheme should be developed for this optical transmission. Students will gain
experience with optoelectronic components and will explore the quality of a transmission
over optical link. This project has a wide range of applications in many implantable
devices in biomedical as well as in RFID systems where it is difficulty to use standard
batteries due to their large sizes. The project requires a very good optimization of both
the primary and the secondary sites to obtain maximum transfer efficiency. A program
that optimizes the electronics components should be carried out using a software
package.
Skills required/developed: Electronics, communication ELEC3240 or ELEC3210
Level of Difficulty: moderate to difficult
Number of Students: 1or 2

YUC-06-5: Wireless Power Generation from Piezoelectric for Microelectronics

Devices
Proposed by: Mehmet Yuce with Reza Moheimani
Description: Piezoelectric devices can be used to generate energy source for low
power microelectronics and wireless systems. They can behave as selfpowered system
and can eliminate the use of batteries and wires. The goal of this project is to build a
prototyping system to harvest energy from piezoelectric via remote controlling. Power
will be generated wirelessly by sending a Radio Frequency (RF) signal to oscillate the
piezoelectric device. An adaptive AC/DC converter circuitry will be designed at
piezoelectric site to achieve an efficient power transfer.
Skills required/developed: Considering the difficulty level of the project, students who
expect to graduate either with Honors I, or II may only apply. ELEC3240 or ELEC3210
Level of Difficulty: Moderate to difficult
Number of Students: 1 or 2

YUC-06-6: A Body Sensor Network

Proposed by: Mehmet Yuce and Jamil Khan
Description: The project aims to implement a wireless medical sensor network
prototyping for biomedical applications for short communication ranges of around 1-2
meters. The emphasis should be given to the low power consumption since the date
transmission rate is low. The work will investigate on the hardware and software
requirements of transmitting physiological signals over short ranges. The system will be
able to able to interface with a variety of medical sensors (such as temperature,
pressure, ECG etc.,).
Skills required/developed: ELEC3240 or ELEC3210, ELEC3500 ELEC2400
Level of Difficulty: Moderate to difficult
Number of Students: 1 or 2

YUC-06-7: Design of a multi-channel analog neural recording device

Proposed by: Mehmet Yuce
Description: Neural Recording experiments are used widely in biomedical engineering,
especially in brain research. A neural recording device is used to experimentally monitor
and stimulate neural arrays. The goal of this project is to design an analog front-end
from discrete components so as to detect small amplitude and noisy neural signal. The
setup for this project should perform data processing and should provide control circuitry
for neural signals.
Skills required/developed: ELEC2400
Level of Difficulty: Moderate to difficult
Number of Students: 1 or 2

YUC-06-8: Jitter Analysis of ADCs for Software Radios

Proposed: Mehmet R. Yuce
Description: Future communication transceivers will provide worldwide access,
coverage, and roaming incorporating with satellites to serve most of population. As to
achieve unity among all standards, tremendous efforts are being devoted to the
development of radios that support multiple standards. A multistandard radio should
have enough flexibility and programmability to be applied to different standards. The
concept of software-defined radio (SDR) is one of emerging technologies enabling the
development of such multistandart systems, reconfigurable and adaptable by software in
the digital domain. Unlike traditional narrow band radio architectures, a software radio
employs an analog-to-digital converter (ADC) that is exposed to a large number of
carriers and channel bandwidths. This project will explore the jitter requirement of high
speed ADCs when applied to mutlticarrier signals in software radios. Currently ADCs are
designed for narrow band signal. It is important to evaluate the role of ADCs for future
wideband radios working with multi carrier signal. The developed technique should be
verified with hardware implementation.
Skills required/developed: ELEC2500 ELEC3240 or ELEC3210
Level of Difficulty: Moderate
Number of Students: 1
 Mehmet Yuce & Gregg Suanning
The following projects are proposed by Gregg Suanning from the School of
Engineering and have been offered to Mechatronics students also. These projects
will be supervised by Mehmet Yuce in the School of Electrical Engineering and
Computer Science. Students wishing to do any of these projects should contact
Mehmet or Gregg for further information.

YUC-06-9: Micro-antenna characterization

Description: The delivery of power and data to a body-implanted medical device is an
important aspect of implant design. In this project, micro-antennas shall be tested and
characterized for their ability to send and receive data and power.

YUC-06-10: Cochlear implant simulator for music and single word comprehension
assessment
Description: The project involves constructing a circuit of several filter banks that will
take input from a microphone, extract particular features of the sound (sound
processing), and convey the processed sound in the form of 22 discrete frequencies.
The completed device can be used in experimentation on normal hearing individuals to
assist in assessing sound-processing techniques.

YUC-06-11: Micro-electrode fabrication using laser ablation

Description: To restore sight to the blind and movement to the paralyzed, appropriate
electrode-tissue interfaces must be devised. This project involves the use of a Nd:YAG
laser to ablate thin foils of metals in order to form biological electrodes. The electrodes
are then to be tested and characterized as part of the project.

YUC-06-12: Support Chip for Biomedical Implants

Description: Many biomedical implants behave according to instructions provided by
external sources, often by radio telemetry. Because of this slave behaviour, the devices
often do not possess the smarts that would allow for a wide-range of measurements
and diagnostics to take place, and be used for worth-while purposes such as patient
monitoring or device control. The purpose of the project is to use a micro-controller to
investigate what can be achieved and how this technology can be applied to future
biomedical implants.

YUC-06-13: Welding of alumina oxide ceramic using laser for medical implants
Description: This project shall investigate the feasibility of welding the alumina oxide
itself (by-passing the need for the thin layers of titanium to be applied as it is currently
done).

YUC-06-14: Syringe Mixing apparatus for Implantable Silicone Mixing

Description: The project will include the design, fabrication and testing of a stepper-
motor controlled syringe mixer to accomplish the task illustrated above. The device
should be micro-controller controlled with mixing parameters (cycles, speed, volume etc)
all readily programmable with buttons etc.

YUC-06-15: Talking scientific calculator with graphics

Description: Visually impaired young people who attend mainstream schools often
require special tools. One tool that is missing is a calculator that co do graphing. This
project aims to construct a calculator that has graphical capabilities and all the
functionality of a standard scientific calculator but with the addition of auditory feedback
fore ach key pressed.

Peter Turner

TUR-06-1: Data Exchange

Proposed By: Peter Turner
Description: This project is based around hardware that has been constructed to allow
Video, Digital and Voice data to be multiplexed onto and de-multiplexed from a single
high-speed synchronous serial link. The project is based on a European standard E1
communications link; however, the data rate is deliberately reduced from the standard
2.048Mbps to 256kbps to accommodate the limited capabilities of the computer
hardware available.
As the hardware for this project has already been designed, this project will involve
writing software (assembler language or C code for the Mitsubishi M16C C) and
completing digital designs (Altera EPLDs) to interface with the hardware to perform the
required communication functions.
Skills required/developed: ELEC3500 ELEC2700
Level of Difficulty: Moderate to Challenging
Number of Students: Suitable for up to three students working as a group.

TUR-06-2: Home Networking

Proposed by: Peter Turner
Description: This project is to implement a wireless Data Acquisition (DAQ) System
using Bluetooth as the wireless technology. Mitsubishi M16C Cs fitted with Ericsson
Bluetooth modules will be used to establish the Bluetooth network. One of these M16C
Cs will act as the DAQ module. This module will need to be programmed and a
significant amount of electronic design will be required to design a Switch-mode power
supply, capable of maintaining an output during a brown-out, as well as interface
electronics that provide significant protection to the C.
The second M16C C will perform two functions and be referred o as the Wireless DAQ
hub. It will gather the data collected from the DAQ module and it will provide a
mechanism for a standard phone to query the status of any DAQs it can communicate
with. This will involve having some form of voice generation software and circuitry
associated with this C.
The data gathered by Wireless DAQ hub will also be relayed to a Bluetooth enabled PC
where a GUI will display the status of any attached DAQ modules. An additional task
may be to implement some form of programmable control functionality with the Bluetooth
enabled PC, e.g. implement a basic graphical programming environment similar to a
simple PLC Ladder Logic application.
Skills required/developed: ELEC3500 ELEC2700
Level of Difficulty: Moderate to Challenging
Number of Students: Suitable for up to three students working as a group.

TUR-06-3: Robot Dog Soccer

Proposed By: Peter Turner
Description: The object of this project is to use 2 Robot Dogs to compete in a game of
Robot Dog Soccer. Each Robot Dog will have a uniquely coloured directional shape
mounted on its back and head so that an overhead USB Camera can be used to
determine its position on the field. Each team will have its own PC and USB Camera.
The PC will receive and interpret the captured images, then send commands via a Direct
Sequence Spread Spectrum (DSSS) wireless communication channel to the Robot
Dogs.
The Robot Dogs have been constructed using Megarobotics AI601 modules. These
modules contain a C with a serial port and circuitry to monitor and position an internal
motor. This motor is positioned by sending commands to the AI601s C via the serial
port. The serial ports of all the AI601 modules are interconnected in a multi-dropped
fashion back to a Mitsubishi M16C C, which is mounted on the back of the Robot Dog.
This M16C C interfaces to an RFWaves DSSS module which allows it to communicate
wirelessly in the 2.4 GHz to 2.5GHz ISM band at data rates up to 1Mbps.
This project was offered last year. All the results will be made available to groups
undertaking this project this year and it is expected that a significant improvement be
made to the Robot Dog Soccer teams performance.
Skills required/developed: ELEC2700 ELEC3500
Level of Difficulty: Moderate to Challenging
Number of Students: Suitable for up to three students working as a group

TUR-06-4: Walking Humanoid Robot

Proposed By: Peter Turner
Description: The aim of this project is to make a humanoid robot navigate through an
obstacle course. To make this difficult task a little easier, this obstacle course will consist
of clearly identifiable objects, i.e. the floor will be green, the walls will be yellow, a rock
will be red, etc.
A CMOS image sensor will be mounted on the humanoid robots head. Images from this
camera will be captured by a local Mitsubishi M16C C and then transmitted via a Direct
Sequence Spread Spectrum (DSSS) wireless communication channel to a PC where the
image processing will be performed. There will be two operating modes for the
humanoid walking robot, these are:
1. Manual in manual mode an operator will be able to guide the robot
through the obstacle course by issuing commands from the GUI on the PC. The operator
will only be able to use the image displayed on the GUI to help navigate the robot.
2. Automatic in automatic mode the PC will implement an image
detection algorithm such that the robot can navigate the obstacle course without any
intervention from an operator.
The Humanoid Robot has been constructed using Megarobotics AI601 modules. These
modules contain a C with a serial port and circuitry to monitor and position an internal
motor. This motor is positioned by sending commands to the AI601s C via the serial
port. The serial ports of all the AI601 modules are interconnected in a multi-dropped
fashion back to a Mitsubishi M16C C, which is mounted on the back of the Humanoid
Robot. This M16C C interfaces to an RFWaves DSSS module which allows it to
communicate wirelessly in the 2.4 GHz to 2.5GHz ISM band at data rates up to 1Mbps.
Two accelerometers will be mounted to the frame of the Humanoid Robot. These
accelerometers will be connected to the frame of the robot and each have 2 analog
outputs which. The outputs of these accelerometers will be a critical factor in the
implementation of the algorithm to monitor the walking robots centre of gravity and keep
the robot upright and balanced.
This project was offered last year. All the results will be made available to groups
undertaking this project this year and it is expected that a significant improvement be
made to the Walking Humanoid Robots performance.
Skills required/developed: ELEC2700 ELEC3500
Level of Difficulty: Hard to Advanced
Number of Students: Suitable for up to three students working as a group

Connell Wagner PPI

CWPPI-06-1: Isolated Signal Conditioning System for DSP applications

Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: The project is the development of a signal conditioning system to be used
with a standard SSH data acquisition card. The project would involve the design and
construction of the hardware, development of the PCB, design of software for the control
of the interface and testing. The objective is to develop a system that has low noise and
is robust enough to be used by CWPPI in routine testing tasks. This is the first stage that
will lead to the design of a 200MSample module that will need to work at voltages up to
1MV for application within CWPPIs high voltage laboratory.
Specifications
24 channel
>= 16 bit resolution
Isolated to 2kV
6 inputs capable of 1kVAC continuous inputs
100kHz bandwidth
Adjustable antialiasing filters
CMRR >140dB DC, >100dB 50Hz, >90db 10kHz and >60dB 50kHz
USB control of gain, offset, AAF frequency
Outcomes: Working prototype of at least 2 channels.
Level of Difficulty: Difficult

CWPPI-06-2: Transmission line temperature measurement

Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: This project requires the construction of a probe that can be placed onto
an in service transmission line of up to 330kV to determine the conductor temperature to
better than 1oC. The project is needed as thermography of transmission lines from
ground level or from helicopters can result in large temperature errors. These errors
substantially effect the ratings of the transmission lines, which can lead to fires or danger
to the public.
The project requires the complete design of the sensor, delivery method and data
communication system. It is envisaged that the sensor will be put on with the aid of a
high voltage operating rod which means that the sensor must be light and rugged. Other
methods may also suit so a few options should be considered such as launching the
device from a launcher and having a mechanism that released the probe by a release
cord or radio controlled mechanism.
Outcomes: Working prototype.
Level of Difficulty: Challenging

CWPPI-06-3: High Dynamic Performance Frequency Transducer

Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: The project is to make a high resolution frequency transducer and PPL
which can accept a variety of input waveforms and produce an analogue output that is a
precise measure of the frequency. An example is that of a once per revolution 50Hz
position encoder signal. In such a case the unit should be able to resolve the frequency
to better than 0.01% of the input frequency. There are of course complications such as
noise, jitter in the input signal and variations in the signal with speed. Depending on how
the project is tackled it could be either done as a hardware implementation or it could be
done with a DSP. The DSP approach allows for many more adjustable elements and
advanced filtering of the input such as with Kalman filtering of the input signal. The
dynamic performance of the system is important and must be well understood with
supporting simulations and tests.
Outcomes: Working prototype.
Level of Difficulty: Difficult to Challenging depending on approach

CWPPI-06-4: Precision Multi-phase Test Source

Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: The project is to develop a multiphase computer controlled source that can
produce precision voltages and currents. These outputs will be used to test the response
of protection relays or calibrate power meters. In this regard the source will have to be
able to produce 4 voltages of up to 240VAC 60VA and 6 currents of up to 15A each. The
controlling source is expected to be an isolated multiphase on-board arbitrary waveform
generator. This should allow waveforms to be phase shifted to a precision of better than
0.1 deg at 50Hz. Standard waveforms at 50Hz should have no material phase error
relative to 120degree separation of phases. Amplitude of all signals should be able to be
varied independently.
The accuracy of waveforms provided as a three phase power set (3V + 3I) should be
<0.03% and should be able to stay within this calibration for a year. The sets of three
waveforms such as voltage and currents should be isolated from each other to at least
500VAC. The arbitrary wave function should have a bandwidth of 1.5kHz.
As this is a challenging project, the power supplies will not be considered as part of the
project.
Outcomes: Working isolated multiphase arbitrary waveform generator with software for
control. Precision voltage and current amplifiers.
Level of Difficulty: Challenging
CWPPI-06-5: Electric Core Imperfection Detector
Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: The cores of large electric motors and generators are subject to damage
and failure. An instrument known as an Electric Core Imperfection Tester (EL CID) can
detect this damage. The instrument is a phase sensitive comparator that measures in-
phase losses to a reference. The objective of this project is to develop a modern version
of the instrument that will allow the user to perform the test and to stored the results in a
manner that fits in with the stop start nature of the testing process. The final outcome is
a contour plot highlighting areas of excessive loss. It is envisaged that the unit will be put
together as much as possible using off the shelf modules such as CPU core and USB
controllers from Z-World Electronics and programmed up in C.
There will be analogue and digital design as well as fabrication work to make the
encoder and control system. The scope of the project is very wide and the use of off-the-
shelf modules will help to ensure that the prototype can be completed during the year.
Outcomes: Working prototype.
Level of Difficulty: Moderate Difficult

CWPPI-06-6: 3 phase transformer ratiometer

Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: The ratios of large power transformers have to be measured in the field to
an accuracy better than 0.1%. This measurement has to be often made in the presence
of a considerable amount of noise as well as standing voltages as high 20VAC (sourcing
up to 100mA). This project is to build a unit capable of measuring 3 phase or single
phase transformer ratios for normal and phase shifting transformers.
This will require the application of DSP principles to deal with the noise scenario and the
design and manufacture of hardware.
Outcomes: Working prototype.
Level of Difficulty: Difficult

CWPPI-06-7: Variable Voltage Variable Frequency Source Control System

Conversion
Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: CWPPI has a 200kVA 415VAC single-phase output VVVF source that is
controlled by an obsolete DEC Alpha computer system. The project is to reengineer the
control system so that it uses currently available technology. This requires a complete
control implementation and the interfacing to all existing interfaces from the existing
system. The project is very involved and requires a person with a high degree of aptitude
in electronics systems and design. It is envisaged that a laptop computer using a fibre
optic cable interface will control the test source.
Outcomes: Working prototype replacement control card.
Level of Difficulty: Very Challenging

CWPPI-06-8: High Voltage Cable Partial Discharge Mapping

Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: High voltage cables are subject to degradation over time and as they age
they develop sites where small electrical discharges occur within the insulation that will
eventually lead to failure of the cable. This project is the development of a system that
can plot the location of these discharge sites along the cable so decisions can be made
about the state of the cable. The signals are gathered by clamping on a high frequency
current sensor to one end of the cable and analysing the measured discharge signals. It
is envisaged that the project will use a high-speed digital oscilloscope and the student
would developed software to control the scope and perform the analysis of the partial
discharge profiles in a graphical format for inclusion into reports.
Outcomes: Software that interfaces to a scope and performs the analysis and presents
results.
Level of Difficulty: Low to Moderate

CWPPI-06-9: 30kV Arbitrary DC source

Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
Description: A computer controlled 0-30kV DC, <10mA source able to output its voltage
to a resistive or capacitive load. It should be able to ramp up or follow an arbitrary DC
signal <100Hz bandwidth. Voltage and current to the load should be measured and
values fed back to a computer control console via a fibre optic cable. The PC software to
drive the source and gather measurements would be part of the project.
Outcomes: Prototype
Level of Difficulty: Moderate to Difficult

CWPPI-06-10: Robust Industrial Voltage Amplifier

Proposed By : Joe Tusek, Connell Wagner PPI (yet to be allocated academic supervisor)
There have been many developments in devices and systems for high power
switchmode amplifiers. The project is to develop a 2 kW low noise switchmode amplifier
able to drive a direct short continuously for general purpose industrial amplifiers. CWPPI
has been using audio style amplifiers but they are subject instability and fragility when
used in industrial applications. The project is to design and prototype high power
amplifier that will be stable driving any load up to its protective limits. Bandwidth <3kHz.
Outcomes: Prototype
Level of Difficulty: Difficult

CSIRO

CSIRO-06-1: An intelligent hot water system

Proposed By : Glenn Platt
Description: Electric hot water systems make up a significant load on the local
electricity network. Currently, such systems are controlled in a relatively simple manner,
based on technology many decades old. Recent changes in the operation of electricity
markets, and the growing interest in demand and peak reduction efforts in electricity
networks mean that there is now an opportunity to improve the control of these systems.
This project involves the design and implementation of an intelligent hot water system
controller. Such a controller may consider factors such as house occupancy and usage
patterns, thermal responses, electricity market prices, and user preferences in deciding
how to operate the hot water heaters.
Skills required/developed: The project would have a hardware component- choosing
and implementing appropriate sensing and control hardware, whilst also involving a
major software component in the design and implementation of intelligent, adaptive
software.
Level of Difficulty: The project is of a moderate difficulty.

CSIRO-06-2: A low-power medium access control technique

Proposed By : Glenn Platt
Description: Medium access control (MAC) techniques are a large area of wireless
communications research. Much research into MAC techniques focuses on issues such
as time of response, error resilience and packet loss that are important to modern
multimedia communications systems. This project is focused on the design and
implementation of a MAC technique that focuses on a different aspect- the energy
consumption of the particular MAC technique. The aim of the project is the design of a
robust MAC technique with as low as possible energy consumption, targetted towards
application in computationally and power constrained hardware such as that found in
wireless sensor networks. It is anticipated the project be simulation based, analysing the
performance of the MAC technique through event based simulation tools.
Skills required/developed: The project would suit a student interested in the design
and analysis of state of the art communications protocols.
Level of Difficulty: Challenging difficulty.

CSIRO-06-3: Optimal management of a grid-connected renewable energy

generation
system
Proposed By : Glenn Platt
Description: CSIRO has a wind turbine and tracking photovoltaic ("solar") generation
system installed at its Mayfield Energy Centre. This system is connected to the electricity
grid and a DC battery bank through a sophisticated inverter/charger system. Optimal
management of such a system is quite complex- for example, a particular user may
choose to maximise profit by selling electricity to the grid, maintaining little reserve
battery storage, whilst another may prefer to maintain maximum battery storage,
preserving the ability to operate in "island" mode during a grid blackout. Many factors are
involved in making such operational decisions, including return on investment,
maintaining the longevity of generation and battery equipment, electrical safety, and
reliability.
This project is based on determining the most optimal management routines for a grid
connected renewable energy generation system. Technical aspects of the project would
focus on designing and implementing an energy monitoring installation to determine the
performance of the management system, whilst the project would also include significant
non-technical aspects such as economic considerations and life cycle analysis.
Level of Difficulty: The project is of a moderate difficulty.

CSIRO-06-4: A "tiny agent" system.

Proposed By : Glenn Platt
Description: Part of CSIRO's intelligent-agent based energy management research
involves the use of state of the art wireless sensor network devices. Wireless sensor
network technology is an exciting area involving the design and application of cheap,
distributable, wireless sensing devices for applications such as environmental
monitoring, battlefield intelligence and condition monitoring. Multi agent systems
research is another growing research area, focused on the analysis, design and
implementation of complex software systems. Such techniques have great potential,
seeing use in applications from energy management to computer generated animation
for movies such as "Lord of the Rings". We believe agent based software techniques
hold great potential, and are interested in their application on wireless sensor network
hardware. Due to the constraints of such hardware, this is a significant challenge. In
2005 CSIRO sponsored a project focused on the introduction of agent based technology
to wireless sensor networks. The project was quite successful, and the "tiny agent"
system that was created shows great promise. We are proposing a continuation of this
project, focusing on the application of more advanced agent algorithms to realise a more
intelligent and autonomous agent system on wireless sensor network devices.
Skills required/developed: The project is based primarily on software design and the
application of advanced computer science techniques.
Level of Difficulty: Difficult

CSIRO-06-5: Development of a microcontroller-based multiplexer for power

monitoring
Proposed By : John Ward
Description: CSIRO has been investigating the use of intelligent control methods to
optimise energy usage. At a residential level this might include air conditioning, hot water
and lighting circuits.
One of the challenges is that power measurements are usually only available as total
site consumption due to the cost of installing monitoring equipment. This project will
develop a system to facilitate low cost monitoring of multiple devices using multiplexed
sensors.
Specific aspects of the project may include:
* Development of multiplexing hardware for sensors
* Design/programming of an embedded microcontroller system, including
communications interface to standard industrial control equipment.
* Analysing the power data to 1) extract power information about
individual equipment items from aggregate data; 2) correlate this
with time and environmental measurements to determine device
efficiency curves.
Level of Difficulty: The project is of a moderate difficulty, extending to challenging if the
determination of efficiency curves is also included.

Neil Weste

Title: Motormouth
Proposed By : Neil Weste
Description: Distraction Osteogenesis is the process of extending a bone by cutting it
and then slowly separating the bone at the right rate to allow bone growth. It is used for
extending chronically broken limbs and in facial reconstruction for jaw lengthening.
Normally, the two ends of the bone are held in place by an external brace that extends
through the skin and has to be manipulated every day to set the daily extension (roughly
1mm a day). This is unsightly and poses an infection risk as there are interior to exterior
passages.
Recently, a small linear motor has been used on leg extensions - the motor being
completely embedded in the leg during distraction and then removed once the bone has
grown.
This project involves assessing the feasibility and design of a motor or actuator small
enough to be used in jaw surgery - i.e. two units must fit in the mouth along the jaw line.
The unit needs to extend roughly 1mm a day for say 10-20 days and then remain
connected until the bone stabilises, at which stage it is remove. The design must include
a mechanism to power or activate the motor or actuator.

Title: Low Power CMOS Receivers for Sensor Radios

Proposed By : Neil Weste
Description: Single chip radio receivers implemented in standard CMOS have a wide
variety of applications. At any frequency, the problem is always to reduce the power
commensurate with battery lifetime requirements.
Radios operating in the ISM bands at 13.5, 27, 40 and 915 MHz can have applications in
environmental sensing.

One application is for a "Bushfire Detector" - a sensor which is distributed in high

volumes in bushland and can detect and then communicate the beginnings of a bush fire
to a remote site.

The objective of this project will be to design a complete receiver for the 915MHz ISM
that can decode narrow-band FSK data in a silicon on sapphire CMOS technology. The
aim of this will be to achieve adequate receiver performance while reducing power to
below that achieved by current implementations. This will require attention to detail at all
levels of design from the system level right down to the circuit level.

If time permits the transmit section will be designed as well.

Access to Cadence IC design tools and Peregrine Semiconductor's CMOS process will
be arranged.
Software Engineering

Huilin Ye
YE-06-1: A feature configuration tool for software product lines
Supervisor: Dr Huilin Ye
Description: There is a growing shortage of software engineers coupled with increased
pressure for shorter time to market, lower costs, higher quality, and more complex
software products. Software product line has emerged as a new development paradigm
that promises to solve these problems via constructive reuse of software assets in a
product family. Modelling feature variability and dependencies will effectively support the
reuse and the derived model will play a central role in configuring the member products
of a product family. This project is intended to develop a feature modeling and
configuration tool.
The tool should have the following functions:
It can build a feature model and present hierarchical and dependency
relationships among the features in a product family.
An existing feature model can be modified, such as introducing new features
and feature dependencies into the model, or removing feature from the models.
Each model must be validated when it is built or updated.
A feature dictionary will be integrated into the tool.
Configuration of member products.
Level of difficulty: Difficult
Number of students: 3

Stephan Chalup
CHA-06-1: Research into Robot Soccer, Rescue, and Assistance
Description: This project area involves several tasks that are investigated within the
Newcastle Robotics Laboratory. Some of them are part of the work of the NUbot robot
soccer team. Other tasks are relevant for current research in the laboratory and aim at
implementing efficient artificial intelligence and machine learning methods on robots.
Work for the competition team may involve travel to Germany in June 2006 and
participation in the competition. Specific topics that can be discussed are, for example:
simulator + camera integration, colour consistency, illumination invariance, localisation,
open behaviour AIBO project, rescNUbot projects.
Level of Difficulty: Advanced and challenging

CHA-06-2: Multi-agent learning with AIBO robots

Description: In 2005 the NUbots developed a simulator to generate strategic soccer
team behaviour for the four legged league. The simulated strategies were transferred
from the simulator to the real robots. In this project new learning paradigms will be
implemented and tested.
Level of Difficulty: Advanced and challenging

Title: Geometric data analysis (individual project or small team)

Description: Using the disciplines high performance computing equipment this project
develops and tests methods for the geometric analysis of high-dimensional data sets.
The methods include dimensionality reduction and associated visualisation techniques.
Level of Difficulty: Advanced and challenging.

Yuqing Lin

Title: Dynamic Web Site Design

Proposed By: Yuqing Lin
Description: Nowadays, a web site is much more than a collection of static web pages.
A dynamic web site might contain one or more of the so called Web Applications where
each application contains many features, for example, retrieving data from a database,
doing some calculation, etc. These features make locating the information easier and
more accurate; they also improve the data presentation and overall, make the web site
more useful.
Many new techniques have been introduced to support richer features or to make the
interaction more efficient. In addition to basic web programming languages such as JSP,
JavaScript and PHP etc., there are many new techniques emerging, such as YAML,
VRML, AJAX, JSON, SPA and AFLAX etc. which could make a web site look better, feel
better and most importantly, improve the interaction.
However, there are many constrains, including bandwidth, DOM etc., limiting the
features which can be implemented in a Web Application, For example, we can not
implement a function which generates large size output on a Web Server side, since it
takes long time to transfer the output to client side and user will close the application
before the results appear on the users computer. So at this stage, the features which
we could implement in the Web Application are not as rich as those in stand alone
software application.
In this project, the students will investigate which of the features used in stand alone
software application are implementable using the currently available techniques. They
will design and implement a web site containing some of the advanced features.
Difficulties: Moderate
Group Size: 1 or 2.
Outcome: The outcome is an advanced feature web site and a research report.

Dr. Kaushik Mahata

Title: Vision and localization of an autonomous car
Proposed by: Dr. Kaushik Mahata
Industry partner: None
Description: Design of intelligent vehicles is one of the leading research areas in
robotics. This project concerns the vision module of an intelligent car. The car is
equipped with a stereo camera system. The video stream captured by the camera
systems is processed in real-time using a digital signal processor (DSP). The following
necessary information are extracted:
1. Number of lanes in the road
2. Relative position of the car with respect to the road
3. Relative position of the car with respect to the other cars in the neighborhood
4. Relative velocity of the car with respect to the other cars in the neighborhood
5. Traffic signal information and road signs
Outcomes expected based on the level of difficulty: This is a new project, and has
not been attempted by any student before. The outcome associated with a difficult level
of difficulty would be preliminary working prototype of the vision module, operating on a
PC. Additional outcome associated to a challenging level of difficulty would be a real
time solution operating on a DSP.
Level of difficulty: Difficult to Challenging.
Background: A good background in signal processing and automatic control will be
advantageous. The software development will need good knowledge in C++/Java.
Number of students: 2

Title: Bio-authentication via iris recognition

Proposed by: Dr. Kaushik Mahata
Description: Authentication using biometric measures is becoming increasingly
common in identification of people. For example, fingerprint is a very common biometric
measure. The goal of this project is to develop an intelligent system, which can identify a
person by examining the iris pattern of his/her eye. The input to the iris recognition
system is the eye-image of a person. This image is used by the system to determine
whether this person is known to the system. The main steps involved are
1. Preliminary image processing
2. Feature extraction
3. Feature matching
The goal of this project is to develop a working prototype which can train itself from a set
of given images, and then can recognize whether a new image belongs to the training
set or not.
Outcomes expected based on the level of difficulty: The outcome associated with a
moderate level of difficulty would be preliminary working solution operating on a PC.
Additional outcome associated to a difficult level of difficulty would be a real time solution
operating on a DSP.
Level of difficulty: Moderate to Difficult.
Background: An exposure to image processing will be advantageous, but not a must.
The software development will need good knowledge in C++/Java.
Number of students: 1

Pablo Moscato
Mos-05-1: Web Data Mining for Functional Genomics Analysis
Proposed by: Alexandre Mendes and Pablo Moscato
Description: Data Mining and its applications to Life Sciences is one of the most
promising areas in Software Engineering. This project will bring data mining and
genomic research together by the development of a search tool to support functional
genomics analysis. The main aim of this project is to develop an automatic functional
genomics analysis tool based on web text mining. A software tool will be developed to
make an automatic search on the IHOP (Information Hyperlinked over Proteins -
http://www.ihop-net.org/UniPub/iHOP/) database using as keywords names of genes of
interest for a given disease. The software will then perform a text mining task looking for
connections between their biological and molecular functions and then present the
results in a graphical fashion. The project will require a good knowledge of Java, Java
Script and XML, as well as graphical packages like the Java AWT or Java Swing. A good
mathematical background is a differential, as the student will need to implement
traditional graph layout techniques.
The student might receive a grant for personal assistance, but the availability and value
of such grant are still under consideration by the University. A definitive answer is due by
the end of January, 2006.
Level of Difficulty: Challenging
Number of Students: 1

Mos-05-2: Cancer Bioinformatics: Clustering, visualization, analysis and ordering

of large datasets
Proposed By: Pablo Moscato
Description: Moscato and collaborators from the Newcastle Bioinformatics Initiative
(NBI) (http://www.cs.newcastle.edu.au/~nbi) has developed a new method for ordering
large datasets based on an Evolutionary Computation method called memetic
algorithm. The method is competitive with other state-of-the-art methods like
EPCLUSTER (provided by the European Bioinformatics Institute
http://ep.ebi.ac.uk/EP/EPCLUST/), CLICK and Expander (provided by Tel-Aviv
University, http://www.cs.tau.ac.il/~rshamir/expander/expander.html), and others listed in
a recent review (see them at http://www.thoracic.org/geneexpress/gene0602.asp)
The objective of this project is to develop a web interface that can help remote users to
upload a microarray dataset (which can be in several different formats) and process
them with the algorithms developed by the NBI members. The system will then return an
ordered dataset, which can be further processed with other methods or web-enabled
tools.
Good programming and teamwork skills are essential, and mathematical background is
highly desirable. A necessary component of this project is a strong initiative of the
students to develop a very useful tool for cancer bioinformatics researchers. This means
that the students should also analyse the functionalities of other web tools and find
creative ways to exploit the integration of those which are available on the public
domain.
Level of Difficulty: Challenging
Number of Students: several

Mos-05-3: Slowing CONCORDE

Proposed By: Pablo Moscato
Description: Moscato and Cotta have proposed in 2003 an evolutionary computation-
based attack of algorithms (A mixed evolutionary-statistical analysis of an algorithm's
complexity. Appl. Math. Lett. 16(1): 41-47 (2003)). In a nutshell, an evolutionary
algorithm tries to evolve instances (inputs) of a given problem, such that the algorithm
(or its software implementation) requires a lot of steps (time) to produce the correct
answer. The idea was tested on the sorting problem, where the evolutionary computation
method has evolved difficult instances for BubbleSort, ShellSort, QuickSort, etc.
CONCORDE (http://www.tsp.gatech.edu/concorde/index.html) is a Traveling Salesman
Problem solver that has provided some of the world records in the exact solution of TSP
instances. However, its running times are only partially correlated with the instance size
http://www.tsp.gatech.edu/concorde/benchmarks/bench99.html. This suggests that it
may be possible to find small, or very small instances for which CONCORDE requires a
lot of CPU time to find the optimal solution. In the project, a computer based interface
should be created to automatically generate (using evolutionary computation or other
metaheuristics) instances that challenge the CONCORDE code. Another desirable
outcome is to provide a grammar-based generation of instances which are all difficult
for CONCORDE, and estimate the rate of growth as a function of the instance size.
Good programming and teamwork skills are essential, C++ programming may also be
necessary. Mathematical background is highly desirable and an interesting in
challenging issues in design of clever data structures and algorithms.
Level of Difficulty: Challenging
Number of Students: 1-2

Mos-05-4: Cross-association to search the biomedical literature

Proposed By: Pablo Moscato
Description: Automatic cross-association is a hot topic in computer science and
software engineering as our society becomes more and more dependent on powerful
methods to analyse the information available on the public domain. As an example,
Chacrkabarti et al, proposed a
method(http://www.cs.cmu.edu/People/deepay/mywww/papers/kdd04.pdf)
that decompose a binary matrix intro disjoint row and column group such that the
rectangular intersections of groups are homogeneous. This method is parameter free
and can help the users to find similarities between certain objects, there are many
applications for which these methods are useful.
In particular, we are interested in developing a prototype system that improves upon
iHOP (http://www.ihop-net.org/UniPub/iHOP/) (a query using gene name `GATA3, would
illustrate the interested reader the functionality of the program/website). We aim at
having a system that would help users to query (in a similar fashion to what iHOP does),
but with a list of genes. An interface should provide a co-citation binary matrix (note that
gene names have also aliases and they should be automatically included). Analysis of
that matrix should be provided to the user (when required) so that suggestions to
analyse the literature can be given.
Good programming and teamwork skills are essential. Mathematical background is
highly desirable and an interesting in challenging issues in design of clever data
structures and algorithms.
Level of Difficulty: Challenging
Number of Students: 2

Regina Berretta
Ber-05-1: Ensemble Classifiers for Analyzing Genomic Data of Heterogeneous
Diseases
Offered By: Pritha Mahata and Regina Berretta
Description: Analyzing genomic data of heterogeneous diseases like breast cancer,
ovarian cancer, etc., is a real challenging task. This project aims to build a classifier
which takes the genomic data for a number of patients, annotated with the subtypes of a
heterogeneous cancer and the data for a new patient as input; and correctly predicts
whether the patient has a recurrent/non-recurrent subtype of the disease. This will help
doctors to decide the required medical treatment for the patient.
Currently available classifiers prove to be very weak for the above task. This project
aims to combine (ensemble) an iterative decision-based multiple-attribute classifier
along with existing ones like C4.5, random forests, Bayesian classifiers, etc. We expect
that this will enhance the classification result for breast cancer and ovarian cancer. The
student will be expected to learn the existing classifiers and implement the ensemble
classifier and test it on the available datasets of breast and ovarian cancers.
Background: Data Structure and Algorithms, Object-Oriented Programming, Skills in
Java.
Number of students: 1
Expected Outcome: A running program implementing the ensemble classifier.
Level of Difficulty: Challenging.

Ber-05-2: WEKA Classification Methods connected to the Alfa-Beta Feature

Selection
Offered By: Regina Berretta
Description: Aside from environmental factors that can trigger some illnesses, we are
more or less likely to have some genetic-based diseases (like for instance some forms of
cancer). Techniques allow Life Science researchers to provide huge volumes of genetic
data. Since each individual's genetic content is unique, but we also share genetic
information, it is possible to trace the relationship between our common and different
pieces of genetic information, linking them with diseases.
One of the problems that the NBI has been working is the Alfa-Beta Feature Selection
Problem. The object is to find a minimal number of genes that can explain a specific
disease. This is a complex combinatorial optimization problem that requires modern
techniques in computer science. The NBI has developed and implemented some
algorithms to solve variants of this problem.
Weka is a collection of machine learning algorithms for data mining tasks, among them
several classification algorithms.The objective of this project is to aggregate several of
the WEKA classification methods to our alfa-beta feature selection implementation.
Experiments using different types of data (different types of cancer, Alzheimer,
Parkinson, etc.) will be necessary.
This project requires good experience in Java programming.
Number of students: 1
Level of Difficulty: Moderate
Ber-05-3: Robust multi-user detection using metaheuristics
Offered By: Regina Berretta and Kaushik Mahata
Description: Multi-user detection (MUD) is a challenging research problem in wireless
and mobile communication where several users are allowed to transmit at the same
time. This is generally achieved by code division multiple access (CDMA) or orthogonal
frequency division multiplexing (OFDM) systems. Traditionally one assumes an additive
white Gaussian noise channel. However, recent research results show that the noise at
the receiver is not Gaussian distributed in a practical scenario. Optimal decoding in non-
Gaussian noise environment is a challenging research problem where we encounter
nonlinear and non-convex optimization problems, which are difficult to solve within the
required time frame. In this project we explore metaheuristics (Genetic Algorithms,
Memetics Algorithms, Tabu Search) for solving these problems.
Outcomes Expected based on Level of Difficulty: Outcome corresponding to the
difficult level would be a working solution operating on PC. Outcome corresponding to
the challenging level of difficulty would be a solution working in real time.
Required Background: Good background in programming, telecommunication and
signal processing will be advantageous.
Number of Students: 1